Homocysteine-Methionine Cycle Is A

HOMOCYSTEINE-METHIONINE CYCLE IS A KEY METABOLIC SENSOR SYSTEM CONTROLLING METHYLATION-REGULATED PATHOLOGICAL SIGNALING - CD40 IS A PROTOTYPIC HOMOCYSTEINE-METHIONINE CYCLE REGULATED MASTER GENE

A Thesis Submitted to the Temple University Graduate Board

In Partial Fulfillment of the Requirements for the Degree MASTER OF SCIENCE

by Chao Gao August 2019

Thesis Approvals:

Dr. Hong Wang, Thesis Advisor, Center for Metabolic Disease Research Dr. Xiao-Feng Yang, co-Advisor, Centers for Inflammation, Translational & Clinical Lung Research, Metabolic Disease Research, and Cardiovascular Research Dr. Jun Yu, Center for Metabolic Disease Research Dr. Stefania Gallucci, Department of Microbiology and Immunology

2019

Chao Gao

ABSTRACT

Homocysteine-Methionine cycle is a key metabolic sensor system controlling

methylation-regulated pathological signaling - CD40 is a prototypic homocysteine-

methionine cycle regulated master

Chao Gao

Master of Science

Temple University, 2019

Thesis Advisory Committee Chair: Hong Wang, MD, PhD

Homocysteine-Methionine (HM) cycle produces a universal methyl group donor

S-adenosylmethionine (SAM), a competitive methylation inhibitor S- adenosylhomocysteine (SAH), and an intermediate amino acid product homocysteine

(Hcy). Elevated plasma levels of Hcy is termed as hyperhomocycteinemia (HHcy) which is an established risk factor for cardiovascular disease (CVD) and neural degenerative disease. We were the first to describe methylation inhibition as a mediating biochemical mechanism for endothelial injury and inflammatory monocyte differentiation in HHcy- related CVD and diabetes. We proposed metabolism-associated danger signal (MADS) recognition as a novel mechanism for metabolic risk factor-induced inflammatory responses, independent from pattern recognition receptor (PRR)-mediated pathogen- associated molecular pattern (PAMP)/danger-associated molecular pattern (DAMP) recognition.

In this study, we examined the relationship of HM cycle gene expression with methylation regulation in human disease. We selected 115 genes in the extended HM

iii cycle, including 31 metabolic enzymes and 84 methyltransferases (MT), examined their mRNA levels in 35 human disease conditions using a set of public databases. We discovered that: 1) HM cycle senses metabolic risk factor and controls SAM/SAH- dependent methylation. 2) Most of metabolic enzymes in HM cycle (8/11) are located in cytosol, while most of the SAM-dependent MTs (61/84) are located in the nucleus, and

Hcy metabolism is absent in the nucleus. 3) 11 up-regulated, 3 down-regulated and 24 differentially regulated SAM/SAH-responsive signal pathways are involved in 7 human disease categories. 4) 8 SAM/SAH-responsive H3/H4 hypomethylation sites are identified in 8 disease conditions.

We conclude that HM cycle is a key metabolic sensor system which mediates receptor-independent MADS recognition and modulates SAM/SAH-dependent methylation in human disease. We propose that HM metabolism takes place in cytosol and that nuclear methylation equilibration requires nuclear-cytosol transfer of SAM, SAH and Hcy.

CD40 is a cell surface molecule which is expressed on antigen presenting cells such as monocyte, macrophage, dendritic cells and neutrophils. The costimulatory pair,

CD40 and CD40L, enhances T cell activation and induce chronic inflammatory disease.

Also, DNA hypomethylation on CD40 promotor induces inflammatory monocyte differentiation in chronic kidney disease. In order to figure out if CD40 is a prototypic

HM cycle regulated master gene, RNA-seq analysis were performed for CD40+ and

CD40- monocytes from mouse peripheral blood and 1,093 differentially expressed genes

(DEGs) were selected from those two groups. All the DEGs modulate as much as 15 functional gene groups such as cytokines, enzymes and transcriptional factors.

Furthermore, CD40+ monocytes activated trained immunity pathways especially in

Acetyl-CoA generation and mevalonate pathway. In HM cycle, CD40 is a prototypic HM cycle regulated master gene to induce the most of the Hcy metabolic enzymes as well as

MT, which can further modulate the methylation-regulated pathological signaling.

The studies in this thesis were funded by the National Institutes of Health (NIH) grant to Dr. Hong Wang.

DEDICATION

To my dearest family

ACKNOWLEDGMENTS

I would primarily like to thank my thesis advisor Dr. Hong Wang at Temple

University Lewis Katz School of Medicine. As a great and dedicative scientist who devoted her life to the research work, she always encouraged me to find out the field I am really interested in and gave me all the freedom to pursue my research. Dr. Wang continuously supported my research and study with her patience, motivation, enthusiasm and immense knowledge. Her guidance helped me in all the time of research and writing of this thesis. Without her great support and encouragement, I cannot go through all the difficulties which I have met during my past two years and finally finish the project and have this thesis. Also, she was always very supportive so I could have opportunities to attend the different international conferences and present my work during past two years.

Secondly, I would like to acknowledge Dr. Xiao-Feng Yang, who is my co- advisor and regularly provided scientific suggestions and insightful comments on my project. His hard questions incented me to widen my research from various perspectives.

He always supported and motivated us to learn new techniques and build collaborations with other colleagues. Dr. Yang’s credible ideas have been great contributors in the completion of the thesis.

Thirdly, I am very grateful to my committee members, Dr. Jun Yu and Dr.

Stefania Gallucci, for their encouragement and positive critiques throughout my committee meetings. Without their passionate participation and input, the project could not have been successfully conducted and finished. I would also like to thank our collaborator Dr. Yan Zhou at Temple Health-Fox Chase Cancer Center for RNA-Seq analysis.

vii

As an international student who first came to the United States after college, I am so delighted to be accepted by the Biomedical Sciences program at Temple University

Lewis Katz School of Medicine. It is really a happy time working with all co-workers and staffs in our school and in Dr. Wang and Dr. Yang’s labs. This lab is a sweet home and all our lab members treated me as their family member. I would like to especially thank

Xiaohua Jiang for her consistent efforts on lab management. I would like to thank Dr. Pu

Fang and Humin Shan for teaching me flow cytometry and single cell suspension. I would like to thank Dr. Hangfei Fu, Ying shao and Remon Cueto for sharing me the strategies on database mining and RNA-seq data analysis. I would also like to thank

Jason Saredy for providing great suggestion on scientific writing.

Last but not least, I would like to thank my lovely family. The reason why I have the courage to go abroad and pursue further study is their huge encouragement. They always told me to have a broad view and experience during my life and never give up my dream. Without their support, I could not have achieved such an accomplishment.

Thanks all the people who have helped me during my study!

viii

TABLE OF CONTENTS

PAGE

ABSTRACT ...... iii

DEDICATION ...... vi

ACKNOWLEDGMENTS ...... vii

LIST OF FIGURES ...... xi

LIST OF TABLES ...... xii

LIST OF ABBREVIATIONS ...... xiii

CHAPTER

1. INTRODUCTION ...... 1

1.1 Overview of Homocysteine and Hyperhomocysteinemia ...... 1

1.2 Homocysteine-Methionine (HM) cycle ...... 1

1.3 Methylation ...... 2

1.3.1 DNA methylation ...... 4

1.3.2 Protein methylation ...... 4

1.3.3 RNA methylation ...... 5

1.4 Metabolism-associated danger signal (MADS) recognition ...... 6

1.5 CD40 and CD40L ...... 6

1.6 Hypothesis of Thesis ...... 7

2. MATERIALS AND METHODS ...... 9

2.1 Selection of extended Homocysteine-Methionine (HM) cycle genes ...... 9

2.2 Subcellular localization of extended HM cycle enzymes ...... 9

2.3 Expression profiles of 115 genes in extended HM cycle in 35 pathological

conditions ...... 9

2.4 Ingenuity Pathway Analysis ...... 10

2.5 Fluorescence-activated Cell Sorting (FACS) and RNA Isolation for RNA-Seq

...... 11

3. RESULTS I ...... 12

3.1 HM cycle senses metabolic risk factors and modulates SAM/SAH-dependent

methylation and pathogenic signaling ...... 12

3.2 All HM cycle enzymes are located in the cytosol, the majority of SAM/SAH-

dependent MT are located in the nucleus ...... 23

3.3 Identification of HM cycle gene expression changes and SAM/SAH-responsive

signal pathways in seven human disease categories ...... 25

3.4 Thirty-three HM cycle-regulated and SAM/SAH-dependent H3/H4 histone

methylations are changed in human disease ...... 38

4. RESULTS II...... 44

4.1 RNA-Seq analysis for mouse CD40+ and CD40- monocyte (MC) ...... 44

4.2 Differentially expressed genes in CD40+ vs CD40- MC ...... 46

4.3 Differentially expressed genes in HM cycle and trained immunity pathways

...... 51

5. DISCUSSION ...... 54

REFERENCES CITED ...... 58

LIST OF FIGURES

Figure 1: HM cycle comprises three major metabolic pathways including

demethylation, remethylation and transsulfuration pathway...... 2

Figure 2: Extended Homocysteine-Methionine (HM) cycle ...... 16

Figure 3: Classical model of receptor-dependent pathogenic signaling ...... 17

Figure 4: Novel receptor-independent MADS recognition contributes to methylation-

regulated pathogenic signaling ...... 17

Figure 5: Cellular localization of HM cycle enzymes and proposed metabolite (SAM,

SAH and Hcy) nuclear-cytosol transfer hypothesis ...... 24

Figure 6: Genes and pathways changed in 7 disease condition categories (A to G) ...33

Figure 7: HM cycle-regulated and SAM/SAH-dependent H3/H4 histone methylations

are changed in human disease ...... 41

Figure 8: Histone MT regulation in 29 disease conditions ...... 42

Figure 9: Sorting strategy for CD40+ and CD40- MC ...... 45

Figure 10: PCA plot results...... 47

Figure 11: Differentially regulated 15 gene functional groups in CD40+ MC ...... 49

Figure 12: Volcano plot showing log2(Fold change, FC) and –log10 (P value) of

24,341 genes ...... 49

Figure 13: Heat map result ...... 50

LIST OF TABLES

Table 1: Extended HM cycle metabolic gene list...... 18

Table 2: Targets and functional implication of 36 SAM/SAH-dependent HMT in

pathophysiological condition...... 21

Table 3: Summary of 35 human diseases in 7 disease categories and their GEO

database...... 26

Table 4: Classification of SAM/SAH-responsive signal pathways in human disease

category ...... 37

Table 5: Summary of 24,341 annotated genes ...... 47

Table 6: 15 functional groups of DEGs ...... 48

Table 7: CD40+ MC gene changes in HM cycle ...... 52

Table 8: CD40+ MC gene changes in trained immunity pathways ...... 53

xii

LIST OF ABBREVIATIONS

AA Amino acid

AD Alzheimer’s disease

Ade Adenosine

AOD Aortic occlusive disease

APC Antigen presenting cell

Arg (R) Arginine

AS Atherosclerosis

BA Breast adenocarcinoma

BC Bladder cancer

βC β cell

CAD Coronary artery disease

CBS Cystathionine-β-synthase

CC Colon Cancer

CCRCC Clear cell renal cell carcinoma

CKD Chronic kidney disease

CRC Colorectal cancer

CTH Cystathionine--lyase

CVD Cardiovascular disease

Cys Cysteine

DAMP Damage-associated molecular pattern

DEG Differentially expressed gene

DNMT DNA methyltransferase

xiii

EC Endothelial cell

ESCC Esophageal Squamous Cell Carcinomas

FACS Fluorescence-activated cell sorting

FC Fold change

FDR False discovery rate

FPKM Fragments per kilobase of transcript per million mapped reads

GA Gastric Adenocarcinoma

GEO Gene Expression Omnibus

Glu Glucose

Hcy Homocysteine

HFH Heterozygote family hypercholesterolemia

HG Hyperglycemia

HGPS Hutchinson-Gilford progeria syndrome

HHcy Hyperhomocysteinemia

HL Hyperlipidemia

HM cycle Homocysteine-methionine cycle

HMT Histone methyltransferase

IC Intrahepatic Cholangiocarcinoma

IL Interleukin

IPA Ingenuity pathway analysis

IS Ischemic stroke

Lys (K) Lysine

LPC Lysophosphatidylcholine

xiv

LPS lipopolysaccharide

MADS Metabolism-associated danger signal

Met Methionine

Monocyte MC

MT Methyltransferase

NSCLC Non-small cell lung cancer

OC Ovarian carcinoma

Ox-LDL Oxidized low-density lipoprotein

Ox-PAPC Oxidized-1-palmitoyl-2-arachidonoyl-sn-glycero-3-

phosphocholine

OSIMOF Old sepsis induced multiple organ failure

PAMP Pathogen-associated molecular pattern

PCA Principal Component Analysis

PC Phosphatidylcholine

PE Phosphatidylethanolamine

PD Parkinson’s disease

PDA Pancreatic Ductal Adenocarcinom

PRR Pathogen recognition receptor

RA Rheumatoid arthritis

RNA-Seq RNA sequencing

SAH S-adenosylhomocysteine

SAM S-adenosylmethionine

SLE Systemic lupus erythematosus

T2DM Type 2 diabetes mellitus

TC T cell

VB12 Vitamin B 12

VSMC Vascular smooth muscle cell

xvi

CHAPTER 1 INTRODUCTION

1.1 Overview of Homocysteine and Hyperhomocysteinemia

Homocysteine (Hcy) is a sulfur containing non-proteinogenic amino acid (AA).

Increased circulating levels of Hcy (above 15 µmol/L) is indicated as

Hyperhomocysteinemia (HHcy), which is generally recognized as an independent risk factor for coronary, cerebral, and peripheral atherosclerosis [1]. HHcy is associated with endothelial cell injury, blood vessel inflammation, ischemic injury, coronary artery disease, stroke, diabetes and kidney diseases.

1.2 Homocysteine-Methionine (HM) cycle

Homocysteine-Methionine (HM) cycle is critical for numerous biochemical processes including AA metabolism and cellular methylation (Figure 1). The HM cycle comprises three major metabolic pathways including demethylation, remethylation and transsulfuration pathway. In the demethylation pathway, methionine (Met), an essential

AA, is converted to S-adenosylmethionine (SAM), the universal methyl donor. After donating its methyl group for cellular methylation, SAM becomes S- adenosylhomocysteine (SAH), which then becomes homocysteine (Hcy) after losing its adenosine (Ade) in a reversible reaction. Hcy is a sulfur-containing non proteinogenic amino acid and can be metabolized for Met biosynthesis via remethylation pathway and for cysteine (Cys) biosynthesis via transsulfuration pathway [2]. Impairment of remethylation and transsulfuration pathways result in elevated levels of Hcy, a human disease teamed as hyperhomocysteinemia (HHcy), which is an established independent

1 risk factor for cardiovascular disease (CVD) including coronary artery disease and stroke, as well as osteoporosis and dementia [3, 4].

MAT1A Remethylation Met SAM Demethylation THF DMG PE X B12 MTs CH2THF MTR PC CH3-X BHMT

CH3THF Betaine Ade PON1 HT Hcy AHCY SAH MARS Ade CBS Cystathionine Transsulfuration CTH Cys

Figure 1. HM cycle comprises three major metabolic pathways including demethylation, remethylation and transsulfuration pathway. In the demethylation pathway, Met is converted to S-adenosylmethionine (SAM), the universal methyl donor.

SAM becomes S-adenosylhomocysteine (SAH) after donating its methyl group and then becomes Hcy after losing its adenosine in a reversible reaction. Hcy can be metabolized for Met biosynthesis via remethylation pathway or for cysteine (Cys) biosynthesis via transsulfuration pathway

1.3 Methylation

Methylation is an essential chemical modification occurring on a variety of molecules including DNA, RNA, protein (especially histone), phospholipid and other small molecules. The HM cycle provides methyl units for most of the cellular methylation reactions, allowing for gene expression regulation, activation and

2 stabilization of these essential molecules and modulation of their biological functions [5].

We were the first to characterize SAH accumulation-based protein/DNA hypomethylation as the essential biochemical mechanism responsible for HHcy-induced endothelial injury and inflammatory monocyte subset differentiation, which are the key pathological features of CVD, diabetes, chronic kidney disease (CKD) and tissue inflammation [6-9].

Numerous evidences shown that dysregulated methylation is involved in cancer,

CVD, CKD and other metabolic diseases. Global DNA hypomethylation and differential promoter methylation are major characteristics of tumorigenesis. Global hypomethylation of the entire genome in peripheral blood leukocytes is demonstrated as the risk factor of colorectal cancer (CRC) and breast cancer [10, 11]. DNA hypomethylation on the promoter region was linked with increased expression of cyclin D2 and maspin in gastric carcinoma and elevated carbonic anhydrase family gene (MN/CA9) in human renal-cell carcinoma [12-14]. DNA hypermethylation on tumor suppressor genes leads to gene silencing and is defined as the mechanistic basis for anti-methylation therapy [5].

Promoter hypermethylation of tumor-suppressor gene retinoblastoma (RB) gene is associated with retinoblastoma, osteosarcoma and small-cell lung carcinoma [15]. DNA hypermethylation of tumor-suppressor gene cyclin dependent kinase inhibitor 2A

(CDKN2A) is correlated with increased risk of breast cancer [16]. DNA methyltransferase (DNMT) inhibitors 5-aza-cytidine (Vidaza) and 5-aza-2’- deoxycytidine (Dacogen) have been used in myelodysplastic syndrome (MDS) therapy

[17, 18].

1.3.1 DNA methylation

Global DNA hypomethylation in advanced atherosclerotic lesions and in peripheral monocyte is also identified as a major feature of CVD and coronary artery disease (CAD) [19-22]. We were the first to propose hypomethylation as the basic biochemical mechanism for CVD [2] and CKD [9]. We reported that hyperhomocysteinemia (HHcy) selectively inhibits endothelial cell (EC) growth, an onset feature of atherosclerosis, by SAH accumulation-resulted DNMT1 inactivation and by suppression of transcription and DNA hypomethylation on cell cycle dependent element

(CDE) consensus element of cyclin A promoter [23]. We and others discovered that

HHcy, increased SAH and reduced SAM/SAH ratio, indicator of hypomethylation status, are correlated with increased atherosclerosis lesion, inflammatory monocyte/macrophage differentiation, severity of CKD and glomerular injury [24-27].

1.3.2 Protein methylation

Protein methylation affects protein–protein interactions, membrane properties, structure and function of chromatin, and gene expression. Numerous evidences demonstrated altered histone methylation on histone 3 (H3) and histone 4 (H4) impacts on CVD, cancer and neurodegenerative disease by chromatin remodeling and gene expression changes [26]. Twelve methylation sites have been identified on H3 and H4, including H3R2, H3K4, H3R8, H3K9, H3R17, H3K23, H3R26, H3K27, H3K36, H3K79,

H4R3 and H4K20 [28], which is catalyzed by 55 histone methyltransferases [29].

Non-histone protein methylation regulates protein subcellular localization and stability in diverse cellular processes. It was suggested that methylation on carboxy group

4 of the C-terminal residue of a protein increases its hydrophobicity and its ability to associate with the cell membrane [30]. We confirmed this theory and reported that hypomethylation on C terminal CAAX motif of oncoprotein p21ras [31] leads to reduced p21ras membrane association, suppressed downstream signaling, and selectively inhibited cell growth in Hcy-treated endothelial cells (EC) [32]. Protein phosphatase 2A (PP2A), a serine/threonine phosphatase impacting on cell growth and signaling, can be methylated at C-terminal leucine 309 residues in a conserved TPDYFL309 motif [33, 34], which leads to increased phosphatase activity [35]. Protein arginine methylation results in mono- and di-methylation of the guanidine nitrogen atoms of arginine. Arginine can be di- methylated either in a symmetrical or asymmetrical manner (SDMA and

ADMA). SDMA can be formed by protein arginine methyltransferases (PRMT)-5/7 cartelization, and ADMA is catalyzed by PRMT-1/2/3/4/6. PRMT catalyze arginine methylation of cell cycle regulator CDK4/6, pRb, E2F1, cyclin D1, p16, p21, p27 and p53 and results in the destabilization of their relevant complex [36]. PRMT also methylate C-terminal glycine-arginine rich (GAR) domain of DNA repair protein MRE11 and regulates its nuclear compartmentalization [37]. Lysine methylation such as trimethylation of calmodulins at lysine 115 decrease the NAD kinase activation [38].

1.3.3 RNA methylation

RNA methylation is a reversible post-translational modification affects RNA stability, normal development, embryonic stem cell (ESC) differentiation and meiotic progression [39]. RNA methyltransferase like 3 (METTL3) and methyltransferase like 14

(METTL4) have tumor-suppressor role and catalyze N6-methyladenosine (m6A)

5 methylation, which is the most common and abundant methylation modification in RNA and accounts for more than 80% of RNA methylation [40]. Depletion of either METTL3 or METTL4 reduces m6A methylation and increases transcript stability in ESC [39], and enhanced GBM stem(-like) cells (GSCs) growth and promoted tumor progression [41].

Hcy induced NOP2/Sun domain family member 2 (NSun2)-mediated mRNA methylation, which increased ICAM‐1 expression in endothelial cells (EC) [42].

Other molecules such as phospholipid can also be methylated.

Phosphatidylethanolamine (PE) is converted to phosphatidylcholine (PC), a membrane phospholipid in mammalian cells [43], by PE-methyltransferase (PEMT). PE methylation reduces the ratio of PE:PC, which is critical for cell membrane integrity and fluidity.

Increased PE:PC ratio is correlated with a decreased fluidity in cell membrane [44]. PE enhances the rigidity of the bilayer, membrane viscosity of cell membrane. Inhibition of

PE methylation leads to the decreased VLDL secretion [45].

1.4 Metabolism-associated danger signal (MADS) recognition

We recently proposed metabolism-associated danger signal (MADS) recognition as a novel mechanism for metabolic risk factor-induced inflammatory responses [46], which is independent from pattern recognition receptor (PRR)-mediated pathogen- associated molecular pattern (PAMP)/danger-associated molecular pattern (DAMP) recognition [47]. We hypothesize that a metabolic sensor system can response to various metabolic risk factors and mediate downstream pathogenic signaling.

1.5 CD40 and CD40L

Co-stimulation is an essential immune switch system to maintain host immune.

Co-stimulatory molecules, such as CD40, CD80 or CD86, expressed on antigen presenting cells (APC) and are critical for T cell activation [48]. CD40, a tumor necrosis factor receptor superfamily member 5 (TNFRSF5), is a cell surface molecule. As a costimulatory molecule, CD40 proposed proinflammatory role in various disease models.

Costimulatory cascade CD40L-CD40 dyad can enhance immune cell activation and chronic inflammatory disease including multiple sclerosis, atherosclerosis, arthritis, SLE and diabetes [49]. Our previous study demonstrated that in CKD, metabolic risk factor

HHcy can be sensed by a metabolic sensor system, leading to the increased SAH and decreased SAM/SAH ratio. This methylation status change caused the DNA hypomethylation on CD40 promoter which contributes to inflammatory monocyte (MC) differentiation [9]. Meanwhile, CD40+ MC is considered as a stronger pro-inflammatory

MC subset compared to intermediated MC and a reliable biomarker for CKD severity [9].

However, the metabolic sensor for methylation status change is not clear and if CD40 is a prototypic HM cycle regulated master gene still need to be investigated.

1.6 Hypothesis of Thesis

In this study, we investigated the hypothesis that first, HM cycle is a key metabolic sensor signal system to determines SAH/SAM dependent methylation and modulating methylation-regulated pathogenic signaling. Second, CD40 is a prototypic

HM cycle regulated master gene. We examined 115 genes of metabolic enzymes and methyltransferase (MT) in extended HM cycle and their expression level in 35 disease conditions. Our study provided systemic analysis using a large set of public databases and

RNA-sequencing analysis which lead to the development of novel molecule and biochemical models and signal networks for human disease. Our hypothesis will be tested with the following specific aims:

Specific aim 1 – Determine whether HM cycle is a key metabolic sensor signal system to regulate SAH/SAM dependent methylation and modulate methylation-regulated pathogenic signaling.

Study 1.1 – To determine HM cycle gene expression level change in human disease

Study 1.2 – To determine the relationship between HHcy-modulated histone methylation in human disease

Study 1.3 –To identify a new model for methylation-related pathogenic signaling

Specific aim 2 – Determine whether CD40 is a prototypic HM cycle regulated master gene for methylation-related pathogenic signaling.

Study 2.1 – To determine differentially expressed genes in CD40+ monocyte based on

RNA-seq analysis

Study 2.2 – To determine significantly changed genes in different functional group and our research relevant signal pathway

CHAPTER 2 MATERIAL AND METHODS

2.1 Selection of extended Homocysteine-Methionine (HM) cycle genes

We selected 115 genes in extended HM cycle, including 31 metabolic enzymes and 84 SAM/SAH-dependent MT from Human metabolome database

(https://www.hmdb.ca/) as listed in Table 1. These genes were utilized to identify gene expression levels, protein subcellular localization, and to generate expression profile in

35 pathological conditions from public mRNA microarray database (Figure 5).

2.2 Subcellular localization of extended HM cycle enzymes

The subcellular localization of 115 extended HM cycle enzymes were determined utilizing the Human Protein Atlas (https://www.proteinatlas.org/) and compartments subcellular location databases (https://compartments.jensenlab.org/) established by cellular organelle proteomics analysis [50]. Subcellular localization of 21 generally accepted intracellular organelle markers were used as internal control for the justification of the reliability of these databases as we reported previously [51]. Based on the extended

HM cycle enzymes subcellular localization obtained, cytosolic and nucleus models were generated in Figure 5

2.3 Expression profiles of 115 genes in extended HM cycle in 35 pathological conditions

Microarray datasets were collected from the Array Express of European

Bioinformatics Institute (https://www.ebi.ac.uk/arrayexpress), which stores data from

9 high-throughput functional genomics experiments. These data include the information of the expression of 115 enzyme genes in extended HM cycle through experiments submitted directly to Array Express or imported from the NIH-NCBI Gene Expression

Omnibus (GEO) database [52]. These datasets were analyzed utilizing GEO2R from

GEO databases as in our previous publications [51, 53]. Differentially expressed genes were defined as p-value ≤ 0.05 and absolute fold change ≥ 2. Induced gene expression was described as greater than or equal to 2-fold and reduced gene expression as less than, or equal to -2-fold. The GEO datasets utilized were grouped as follows: 1) Metabolic disease: GSE6088, GSE15524, GSE23561, GSE25724, GSE15653; 2) Metabolite treatment: GSE15575, GSE46262, GSE68021, GSE72633, GSE9490, GSE43166; 3)

Vascular disease: GSE9874, GSE57691, GSE23561, GSE22255; 4) Aging disease:

GSE41751, GSE6613, GSE1297, GSE13205; 5) Autoimmune disease: GSE75343,

GSE35958, GSE23561, GSE46907, GSE27335; 6) Digestive cancer: GSE45670,

GSE79973, GSE58561, GSE45001, GSE13471 and 7) Other system cancer: GSE36668,

GSE46602, GSE70947, GSE21933, GSE45184, GSE68417. We generated an association between the gene expression from above datasets and H3/H4 histone methylation changes obtained from literature mining in Figure 7.

2.4 Ingenuity Pathway Analysis

Identified significant differentially expressed extended HM cycle genes from the microarray datasets were analyzed using Ingenuity Pathway Analysis (IPA, http://www.ingenuity.com/) running core analysis as our previous publications [51]. The

10 identified HM cycle-related and SAM/SAH-responsive signal pathways were used to characterize pathophysiological relationship with seven human disease categories.

2.5 Fluorescence-activated Cell Sorting (FACS) and RNA Isolation for RNA-Seq

We obtain CD40+ and CD40- monocytes from mouse peripheral blood. After digesting the red blood cell, samples were washed with FACS buffer (2% FBS in PBS), cells were stained with surface antibody anti-CD40 for 30 minutes at room temperature.

All cell-sorting experiments were performed using an Aria Cell Sorter (BD Biosciences) in Temple University Lewis Katz School of Medicine Flow Cytometry Core. The CD40+ and CD40- monocytes were sorted directly into TRIzol, and RNA was extracted using miRNeasy Mini kit (Qiagen).

Total RNA libraries were prepared by using Pico Input SMARTer Stranded Total

RNA-Seq Kit (Takara). In short, 250 pg-10 ng total RNA from each sample was reverse- transcribed via random priming and reverse transcriptase. Full-length cDNA was obtained with SMART (Switching Mechanism At 5' end of RNA Template) technology.

The template-switching reaction keeps the strand orientation of the RNA. The ribosomal cDNA is hybridized to mammalian-specific R-Probes and then cleaved by ZapR.

Libraries containing Illumina adapter with TruSeq HT indexes were subsequently pooled and loaded to the Hiseq 2500. Single end reads at 50 bp with 40 million reads per sample were generated for bioinformatic analysis.

CHAPTER 3 RESULTS I

3.1 HM cycle senses metabolic risk factors and modulates SAM/SAH-dependent methylation and pathogenic signaling

The extended HM cycle is essential for Hcy metabolism which contains 8 metabolic pathways facilitated by a great many of metabolic enzymes. We dig out 115 genes in the extended HM cycle including 1 Hcy synthesis gene, 14 Hcy metabolism genes, 84 MT genes (4 DNA MT, 24 RNA MT, 30 histone lysine MT, 6 histone arginine

MT, 9 other protein MT and 11 ungrouped MT), and 16 adenosine (Ade) metabolism/transporter genes (Table1 and Figure 2).

In the HM cycle, demethylation, remethylation and transsulfuration are the primary pathways. Firstly, in the demethylation pathway, Met is converted to SAM, which is a universal methyl donor for most of the cellular methylation. After providing its methyl group for cellular methylation, SAM becomes SAH, which is a competitive methylation inhibitor. After losing its adenosyl group, SAH becomes Hcy. Secondly, in the remethylation pathway, Hcy is converted back to Met by receiving a methyl group from folate cycle metabolism or choline/betaine metabolism. Thirdly, in the transsulfuration pathway, Hcy is converted to another sulfur-containing amino acid cysteine (Cys) through intermediate amino acid cystathionine. These top 3 pathways compose of the basic HM cycle which contains 4 major metabolites Met, SAM, SAH and

Hcy and their related metabolic enzymes. Pathway 4 to 8 are secondary pathways.

Pathway 4 describes all SAM/SAH-dependent methylation reactions which are catalyzed by 84 methyltransferases (MT). In pathway 5, folate cycle facilitates Met production. In

12 pathway 6, choline and betaine also contribute methyl group for Met synthesis. Pathway

7 forms Hcy and is the sole source of Hcy synthesis. Hcy can reversibly metabolized to homocysteine thiolactone (HT). Pathway 8 describes Ade transport and metabolism. Ade is transported into cell by a family of adenosine transporter proteins which is called equilibrative nucleoside transporter and produced during Hcy synthesis. Impairment of remethylation and transsufuration pathways, alteration in metabolic enzymes (Met synthase, methyl tetrahydrofolate reductase (MTHFR), cystathionine β-synthase (CBS), and cystathionine-γ-lyase (CγL or CTH) or deficiency in cofactors (vitamin B6 , B12 , folate) is considered as the cause of human metabolic disease HHcy, an established independent risk factor for CVD and other degenerative disease [3, 4, 54].

Classical pathogenic risk factor recognition indicates the ligand-receptor specific molecular recognition. It is a receptor dependent recognition. Infectious pathogens

(pathogen-associated molecular pattern) or noninfectious tissue injury (damage- associated molecular pattern) are considered as pathogenic risk factors which can be recognized by specific receptors. Figure 2 demonstrated the classical model of receptor- dependent pathogenic signaling, which emphasizes ligand-receptor specific molecular recognition. Pathogenic risk factor ligands are defined as pathogen-associated molecular pattern (PAMPs) and damage-associated molecular pattern (DAMPs). Through specific molecular recognition, PAMPs or DAMPs bind to pathogen recognition receptor (PRR) located on the cell-surface or in the cytoplasm or nucleus, which triggers downstream pathogenic signaling response. The PAMP/DAMP-PRR recognition activated signaling leads to post-translational modifications of signaling molecules such as NF-kB and AP-1 activation [55], leading to gene expression changes and disease development [56, 57].

Based on our recent discoveries, we proposed a novel model of "receptor- independent MADS-mediated methylation regulated pathogenic signaling" (Figure 3).

This model describes receptor-independent metabolism-associated danger signal (MADS) recognition and emphasizes metabolic risk factor/sensor-mediated MADS recognition which primarily leads to methylation-dependent pathogenic signaling. We propose that

HM cycle is a metabolic sensor system, which produces SAM and SAH as the universal methyl group for cellular methylation and the competitive MT inhibitor. Histone MT uses

SAM as the methyl donor for histone methylation which causes histone tail structural modification and chromatin epigenetic modification. SAM/SAH-dependent methylation modify methylation on 6 classes of molecules including histone lysine (K) and arginine

(R) residues, other proteins, RNA, DNA and ungrouped small molecules (arsenite, catechol, diphthamide, histamine, nicotinamide, N-acetylserotonin, phosphatidylethanolamine, phenylethanolamine, tryptamine, thiopurine and coenzyme).

We presented 36 SAM/SAH-dependent histone MT (HMT, 30 lysine MTs, 6 arginine MTs), their distinguished modification site and specific function impacts in

Table 1. Recent studies established H3K4 and H3K9 as the major lysine (K) methylation sites on histone [58]. H3K4 and H3K9 can be mono-, di- or tri-methylated. Their respective MTs were found reduced in multiple human cancers [59]. H3K4 methylation is associated with the promoters of actively transcribed genes and impact on transcriptional elongation [59]. H3K9me1 is enriched at the transcriptional start site (TSS) of active genes while H3K9me2/3 are both found at silenced genes and euchromatin/heterochromatin silencing [58, 59]. H3K9 MTs were found induced in gastric adenocarcinoma [60, 61]. H3K27 can be di- or tri-methylated which were

14 associated with gene silencing in general [62]. H3K27me3 has the function for shutting down transcription, H3K27me3 correlates with transcriptional silencing of HoxA genes

[61]. It is reduced in human cancer [59]. H3K36 can be mono-, di- or tri-methylated, which are diversely associated with gene transcription, cell cycle, mRNA splicing and

DNA repair [59]. H3K79 and H4K20 methylation have been studied recently. H3K79 methylation is implicated in several processes, including transcription elongation by

RNA polymerase II, DNA damage response and cell cycle checkpoint activation [63, 64].

H3K79me3 are generally associated with transcriptional activity [61]. H4K20me1 is associated with transcriptional activation. H4K20me2 is important for cell cycle control.

H4K20me3 is associated with repression of transcription when present at promoters [61,

65].

Histone arginine methylation on H3/R2, 8, 17, 26 and H4R3 can be facilitated by various isoforms of PRMT. H3R2me2a, R2 asymmetric demethylation, is a repressive marker for transcription silencing [66, 67]. Whereas, H4R3me2a, R3 asymmetric demethylation, is mainly associated with actively transcribed promoters [68].

H3R8me1/2 is linked to transcriptional repression, while H3R17/R26me1/2 are associated with transcription activation [69, 70].

We also presented 48 SAM/SAN-dependent non-HMT and their functional.

SAM/SAN-dependent non-HMTs facilitate various methylation on other proteins, RNA,

DNA, phospholipids and other biochemical small molecules, and regulates gene expression, molecular stability and biosynthesis, and metabolic status. Dysregulation of non-HMT is associated with cardiovascular disease, cancer, neurodegenerative disease, metabolic disorder and other severe diseases

Figure 2. Extended Homocysteine-Methionine (HM) cycle. HM cycle includes 3 major

pathways. ① In the demethylation pathway, Met is converted to SAM, which becomes

SAH after donating the methyl group for cellular methylation. SAH is then converted to

Hcy, a dual directional reaction, interacting with adenosine metabolism. ② In the remethylation pathway, Hcy is converted back to Met by regaining the methyl group

from folate cycle or choline/betaine metabolism. ③ In the transsulfuration pathway, Hcy

is converted to Cys. The extended HM cycle includes these three pathways and its

⑤ ⑥

interactive metabolisms, including ④ 84 MTs, folate cycle, choline/betaine ⑧ metabolism, ⑦ Hcy synthesis and adenosine metabolism. Abbreviations for gene and enzyme names are explained in Table 1.

Ligands Specific receptors (risk factors) (molecular recognition) Response Pathogen/injury PRR generated molecules DAMPs (Cell-surface Pathogenic + and others PAMPs /Intracellular signaling /Nuclear) Figure 3. Classical model of receptor-dependent pathogenic signaling. This classical model is featured by ligand-receptor specific molecular recognition.

MADS recognition Responses Metabolic sensor (SAM/SAH-dependent methylation)

(Class) substrate Lys-CH3 (I). Histone (K) (VI). Ungrouped small molecules -CH -CH Met SAM Ar 3 Ca 3 Arg-CH MTs (II). Histone (R) 3 -CH3 HHcy, -CH3 Di HM cycle - NAS HG, HL... NAM -CH3 Methylation CH3 -CH -CH (III). Non-hist protein 3 Ha 3 -regulated -CH3 Hcy SAH PE pathogenic -CH3 (IV). RNA -CH3 Tr Th -CH3 signaling -CH3 -CH ? (V). DNA -CH3 PEOH Co 3

Figure 4. Novel receptor-independent MADS recognition contributes to methylation-regulated pathogenic signaling. This model describes receptor- independent MADS recognition and emphasizes metabolic risk factor/sensor-mediated pathogenic signaling and HM cycle-regulated and SAM/SAH-dependent methylation.

Abbreviations: HHcy, hyperhomocysteinemia; HG, hyperglycemia, HL, hyperlipidemia;

Ar, arsenite; Ca, catechol; Di, diphthamide; Ha, histamine; NAM, nicotinamide; NAS, N- acetylserotonin; PE: phosphatidylethanolamine; PEOH, phenylethanolamine; Tr, tryptamine; Th, thiopurine; Co, coenzyme.

Table 1. Extended HM cycle metabolic gene list

Extended HM cycle metabolic gene Abbreviation Full Name S-adenosylhomocysteine hydrolase (1): AHCY Adenosylhomocysteinase Homocysteine metabolic enzymes (14): CBS Cystathionine-beta-synthase CTH Cystathionine gamma-lyase PON1 Paraoxonase 1 MARS Methionyl-tRNA synthetase MTR 5-methyltetrahydrofolate-homocysteine methyltransferase SHMT1 Serine hydroxymethyltransferase 1 SHMT2 Serine hydroxymethyltransferase 2 MTHFR Methylenetetrahydrofolate reductase BHMT Betaine--homocysteine S-methyltransferase MAT1A Methionine adenosyltransferase 1A PLD1 Phospholipase D1 PLD2 Phospholipase D2 CHDH Choline dehydrogenase ALDH7A1 Aldehyde dehydrogenase 7 family member A1 Methyltransferase (MT, 84): Class I- Histone-lysine N-MT (30) ASH1L ASH1 like histone lysine methyltransferase CAMKMT Calmodulin-lysine N-methyltransferase DOT1L DOT1 like histone lysine methyltransferase EHMT1 Euchromatic histone lysine methyltransferase 1 EHMT2 Euchromatic histone lysine methyltransferase 2 EZH1 Enhancer of zeste 1 polycomb repressive complex 2 subunit EZH2 Enhancer of zeste 2 polycomb repressive complex 2 subunit KMT2A Lysine methyltransferase 2A KMT2B Lysine methyltransferase 2B KMT2C Lysine methyltransferase 2C KMT2D Lysine methyltransferase 2D KMT2E Lysine methyltransferase 2E KMT5A Lysine methyltransferase 5A NSD1 Nuclear receptor binding SET domain protein 1 PRDM2 PR/SET domain 2 PRDM6 PR/SET domain 6 PRDM7 PR/SET domain 7 PRDM9 PR/SET domain 9 SETD1A SET domain containing 1A SETD1B SET domain containing 1B SETD2 SET domain containing 2 SETD3 SET domain containing 3 SETD7 SET domain containing 7 18

Abbreviation Full Name SETDB1 SET domain bifurcated 1 SETDB2 SET domain bifurcated 2 SETMAR SET domain and mariner transposase fusion gene SMYD2 SET and MYND domain containing 2 SMYD3 SET and MYND domain containing 3 SUV39H1 Suppressor of variegation 3-9 homolog 1 SUV39H2 Suppressor of variegation 3-9 homolog 2 Class II- Histone-arginine N-MT (6) CARM1 Coactivator-associated arginine methyltransferase 1 PRMT1 Protein arginine methyltransferase 1 PRMT2 Protein arginine methyltransferase 2 PRMT5 Protein arginine methyltransferase 5 PRMT6 Protein arginine methyltransferase 6 PRMT7 Protein arginine methyltransferase 7 Class III- Other protein MT (9) METTL11B Methyltransferase like 11B GAMT Guanidinoacetate N-methyltransferase GNMT Glycine N-methyltransferase HEMK1 HemK methyltransferase family member 1 LRTOMT Leucine rich transmembrane and O-methyltransferase domain containing ICMT Isoprenylcysteine carboxyl methyltransferase LCMT1 Leucine carboxyl methyltransferase 1 NTMT1 N-terminal Xaa-Pro-Lys N-methyltransferase 1 PCMT1 Protein-L-isoaspartate (D-aspartate) O-methyltransferase Class V- RNA MT (24) ALKBH8 AalkB homolog 8, tRNA methyltransferase CMTR1 Cap methyltransferase 1 COQ3 Coenzyme Q3, methyltransferase DIMT1 DIM1 dimethyladenosine transferase 1 homolog EMG1 EMG1, N1-specific pseudouridine methyltransferase FTSJ1 FtsJ RNA methyltransferase homolog 1 FTSJ3 FtsJ RNA methyltransferase homolog 3 HENMT1 HEN methyltransferase 1 MRM2 Mitochondrial rRNA methyltransferase 2 NSUN2 NOP2/Sun RNA methyltransferase family member 2 RNMT RNA guanine-7 methyltransferase TGS1 Ttrimethylguanosine synthase 1 TRDMT1 tRNA aspartic acid methyltransferase 1 TRMT1 tRNA methyltransferase 1 TRMT13 tRNA methyltransferase 13 TRMT2B tRNA methyltransferase 2 homolog B TRMT44 tRNA methyltransferase 44 homolog TRMT5 tRNA methyltransferase 5 TRMT61A tRNA methyltransferase 61A 19

Abbreviation Full Name TRMT61B tRNA methyltransferase 61B METTL1 Methyltransferase like 1 METTL2B Methyltransferase like 2B METTL3 Methyltransferase like 3 METTL6 Methyltransferase like 6 Class IV- DNA MT (4) DNMT1 DNA methyltransferase 1 DNMT3A DNA methyltransferase 3 alpha DNMT3B DNA methyltransferase 3 beta DNMT3L DNA methyltransferase 3 like Class VII- Ungrouped MT (11) AS3MT Arsenite methyltransferase ASMT Acetylserotonin O-methyltransferase COMT Catechol-O-methyltransferase COQ5 Coenzyme Q5, methyltransferase DPH5 Diphthamide biosynthesis 5 HNMT Histamine N-methyltransferase INMT Indolethylamine N-methyltransferase NNMT Nicotinamide N-methyltransferase PEMT Phosphatidylethanolamine N-methyltransferase PNMT Phenylethanolamine N-methyltransferase TPMT Thiopurine S-methyltransferase Adenosine metabolic enzymes (16): SLC28A1 Solute carrier family 28 member 1 SLC28A2 Solute carrier family 28 member 2 SLC28A3 Solute carrier family 28 member 3 SLC29A1 Solute carrier family 29 member 1 SLC29A2 Solute carrier family 29 member 2 SLC29A3 Solute carrier family 29 member 3 SLC29A4 Solute carrier family 29 member 4 NT5C1A 5'-nucleotidase, cytosolic IA NT5C1B 5'-nucleotidase, cytosolic IB NT5C2 5'-nucleotidase, cytosolic II NT5C3 5'-nucleotidase, cytosolic III NT5C 5', 3'-nucleotidase, cytosolic NT5E 5'-nucleotidase ecto NT5M 5',3'-nucleotidase, mitochondrial ADA Adenosine deaminase ADK Adenosine kinase

Table 1. Extended HM cycle gene list. 115 genes in extended HM cycle are divided into four groups: 1) S-adenosylhomocysteine hydrolase metabolism (1 gene), 2) homocysteine

20 metabolism (14 genes), 3) methyltransferase (84 genes) and 4) adenosine metabolism (16 genes).

Table 2. Targets and functional implication of 36 SAM/SAH-dependent HMT in pathophysiological condition

Gene name Modification site Function Pathophysiology Class I: Histone-lysine MT (30 MTs) SETD7 H3K4me1 ASH1L H3K4me1/3 KMT2A H3K4me1/2/3 KMT2B H3K4me1/2/3 KMT2C H3K4me1/2/3 Transcriptional KMT2D H3K4me1/2/3 activation Reduced in cancers KMT2E H3K4me3 Transcriptional SETD1A H3K4me1/2/3 elongation SETD1B H3K4me1/2/3 SMYD3 H3K4me2/3 PRDM7 H3K4me3 PRDM9 H3K4me3 PRDM2 H3K9me1/2/3 Transcriptional EHMT1 H3K9me1/2 activation EHMT2 H3K9me1/2 Induced in gastric SETDB1 H3K9me2/3 adenocarcinoma SUV39H1 H3K9me2/3 Gene silencing SUV39H2 H3K9me2/3 SETDB2 H3K9me3 EZH1 H3K27me2/3 Silencing HoxA, a Reduced in cancers EZH2 H3K27me2/3 leukemia gene Transcription, H3K36me1/me2, DNA repair, cell SETD3 H3K4* cycle, SETMAR H3K36me1/me2, pre-mRNA Dysregulated in NSD1 H3K9* splicing & breast cancer SMYD2 H3K36me2/3 inappropriate SETD2 H3K36me2 initiation H3K36me3 suppression in coding region Transcription, Induced in breast DOT1L H3K79me1/2/3 development, cell cancer cycle, DNA repair

KMT5A H4K20me1 DNA repair, DNA Reduced in breast replication & cancer, lymphomas & PRDM6 H4K20me3 chromatin colon adenocarcinomas. compaction Body growth and Deficiency in cystinuria, CAMKMT Unknown substrate somatosensory intellectual disabilities & development mitochondrial disease Class II: Histone-arginine MT (6 MTs) PRMT4 H3R26me1/2, Increased in breast cancer, (CARM1) H3R17me1/2 informative prognostic marker Promotes breast cancer cell PRMT1 H4R3me1/2 migration/ invasion mRNA splicing, Inhibits breast cancer cell PRMT2 H3R8 transcription proliferation regulation, signal H3R2me1/2, transduction and PRMT5 H3R8me1/2, Increased in breast cancer stem cell protein-protein H4R3me1/2 interaction H3R2me1/2, PRMT6 Involved in breast cancer H4R3me1/2 H3R2me1/2, PRMT7 Promotes breast cancer metastasis H4R3me1/2

Table 2. Classification of targets and functional implication of 36 SAM/SAN- dependent histone MT in pathophysiological conditions. 36 SAM/SAH-dependent MT are classified into two classes: Histone-lysine MT and Histone-arginine MT. Targets and functional implication of 36 SAM/SAH-dependent HMT in pathophysiological conditions and their classification, function and expression changes are listed. Target site of H-MT are indicated. Abbreviation: H, histone; R, arginine; K, lysine; me1, mono- methylation; me2, di-methylation; me3, tri-methylation; HMT, histone methyltransferase.

Abbreviations for gene and enzyme names are explained in Table 1

3.2 All HM cycle enzymes are located in the cytosol, the majority of SAM/SAH- dependent MT are located in the nucleus

The paradigm of a cytoplasmic HM cycle synthesizing/eliminating metabolites that are transported into/out of the nucleus as required has been challenged by detection of significant nuclear levels of several enzymes of this pathway including BHMT [71].

We hypothesize that the majority of extended HM cycle enzymes are mainly localized in the cytosol and SAM generated in the cytosol is transferred or moved into the nucleus, which serves as the methyl donor for histone methyltransferases. To test this hypothesis, we analyzed the subcellular locations of 115 SAH metabolic enzymes with the immunohistochemical data deposited in the Human Protein Atlas database website

(https://www.proteinatlas.org/). For a few enzymes without the data in the Human Protein

Atlas database, the subcellular location of enzymes was analyzed with the Compartments subcellular location database (https://compartments.jensenlab.org/Search). As we reported previously [51], to demonstrate the reliability of these databases, subcellular localization of 21 generally accepted intracellular organelle markers were analyzed. Our report demonstrated that the Compartments database we used is extremely reliable to predict the subcellular localizations of proteins of interest [50, 51].

We have 5 major findings; 1) All essential HM cycle metabolic enzymes were localized in the cytosol but not nucleus, except for Hcy clearance enzyme CBS, which was identified in both cytosol and nucleus (Figure 5), suggesting that HM cycle is completely operative only in the cytosol and that Hcy clearance is critical in the nucleus.

2) Less (25%, 21/84) SAM/SAH-dependent MT are located in the cytosol. 3) Most (73%,

61/84) SAM/SAH-dependent MT are located in the nucleus. 4) SAM, SAH and Hcy can

23 be transferred between nucleus and cytosol. Because most of the enzymes for SAM synthesis, SAH clearance and Hcy synthesis are missing in the nucleus, we propose the model of cytosol-nucleus SAM, SAH and Hcy transfer. 5) CBS facilitates Hcy clearance in the nucleus. CBS is localized in the nucleus, suggesting that Hcy could be potentially

“toxic” in the nucleus and has to be removed immediately via CBS activity to prevent protein covalent modification by the metabolite of Hcy, Hcy thiolactone, protein N- homocysteinylation. Since Hcy, SAM and SAH are AA and can go throuth the nuclear pore freely, and only Hcy clearance enzyme CBS exists in the nucleus, we propose that

HM metabolism takes place in the cytosol, that nuclear methylation equilibration requires a nuclear-cytosol transfer of SAM/SAH/Hcy, and that Hcy clearance is essential for genetic protection.

Cytosol Nucleus MAT1A Met SAM Met SAM Sub. Sub. MTR BHMT 21/84 MT (25%) 61/84 MT (72.6%)

Sub-CH3 Sub-CH3 MARS HT Hcy SAH HT Hcy AHCY SAH CBS CBS Cys Cys

Figure 5. Cellular localization of HM cycle enzymes and proposed metabolite (SAM,

SAH and Hcy) nuclear-cytosol transfer hypothesis. All essential enzymes in the HM metabolic cycle are located in the cytosol. Only 21 out 84 MT (25%) are located in the cytosol. Some of the HM cycle metabolic processes are not active in the nucleus because their metabolic enzymes are not identified there. These inactive metabolic processes are indicated by dash lines. The enzymes for SAM synthesis, SAH clearance and Hcy

24 synthesis are missing in the nucleus. We propose the model of cytosol-nucleus SAM,

SAH and Hcy transfer.

3.3 Identification of HM cycle gene expression changes and SAM/SAH-responsive signal pathways in seven human disease categories

We examined gene expression of 115 enzymes in extended HM cycle in microarray datasets of 35 human disease conditions in NIH-NCBI GEO database

(https://www.ncbi.nlm.nih.gov/geo/). These 35 disease conditions were classified into 7 disease categories and presented in Table 3: a) 5 metabolic diseases (heterozygote family hypercholesterolemia (HFH), type 2 diabetes mellitus (T2DM), obesity, obesity with

T2DM); b) 6 metabolite treatments (high glucose (450mg/dl), high glucose (25mM), oxidized low-density lipoprotein (Ox-LDL), oxidized 1-palmitoyl-2-arachidonoyl-sn- glycero-3-phosphocholine (Ox-PACP), Hcy (100umol/L), Vitamin B12 deficiency); c) 4 vascular diseases(atherosclerosis (AS), aortic occlusive disease (AOD), coronary artery disease (CAD), ischemic stroke (IS)); d) 4 aging diseases (Hutchinson-Gilford progeria syndrome (HGPS), Parkinson's disease (PD), Alzheimer's disease (AD), old sepsis induced multiple organ failure (OSIMOF)); e)5 autoimmune diseases (psoriasis, osteoarthritis, rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), asthma); f)

5 digestive cancers (esophageal squamous cell carcinomas (ESCC), gastric adenocarcinoma (GA), pancreatic ductal adenocarcinoma (PDA), intrahepatic cholangiocarcinoma (IC), colon cancer (CC)); and g) 6 other cancers (ovarian carcinoma

(OC), prostate cancer (PC), breast adenocarcinoma (BA), non-small cell lung cancer

(NSCLC), bladder cancer (BC), clear cell renal cell carcinoma (CCRCC)). Genes with

25 significantly changed mRNA levels (absolute folder change ≥ 2, p ≤ 0.05) were selected and further analyzed for signal pathway using Ingenuity Pathway Analysis (IPA, http://www.ingenuity.com/), which was termed as SAM/SAH-responsive signal pathways.

The identified significantly changed genes and their corresponsive SAM/SAH-responsive signal pathways are presented in Figure 6 A-G.

Table 3 Summary of 35 human diseases in 7 disease categories and their GEO database

Disease category Disease GEO database

GSE6088, GSE15524, Metabolic Heterozygot Family Hypercholesterolemia, Obese, GSE23561, GSE25724, diseases T2DM (T cell), T2DM (β cell), Obese-T2DM GSE15653 GSE15575, High glucose (450mg/dl), High glucose (25mM), Metabolite GSE46262,GSE68021, Ox-LDL, Ox-PAPC, Hcy (100umol/L), treatments GSE72633, GSE9490, Vitamin B12 deficiency GSE43166 Atherosclerosis, Aortic occlusive disease, Coronary GSE9874, GSE57691 Vascular diseases artery disease, Ischemic stroke GSE23561, GSE22255 Hutchinson-Gilford progeria syndrome, Parkinson, GSE41751, GSE6613, Aging diseases Alzheimer, Old sepsis induced multiple organ failure GSE1297, GSE13205 GSE75343, GSE35958 Autoimmune Psoriasis, Osteoarthritis, Rheumatoid arthritis, SLE, GSE23561, GSE4690, diseases Asthma GSE27335 Esophageal squamous cell carcinomas, Gastric GSE45670, GSE79973, Digestive cancers adenocarcinoma, Pancreatic ductal adenocarcinoma, GSE58561, GSE45001, Intrahepatic cholangiocarcinoma, Colon cancer GSE13471 Ovarian carcinoma, Prostate cancer, Breast GSE36668, GSE46602, Other cancers adenocarcinoma, Non-small cell lung cancer, Bladder GSE70947, GSE21933, cancer, Clear cell renal cell carcinoma GSE45184, GSE68417

Table 3 Summary of 35 human diseases in 7 disease categories and their GEO database. 35 disease conditions were divided into 7 disease categories. In each disease category, specific disease name and their corresponding microarray dataset were listed above.

In category A (5 metabolic diseases), we found that 19 out of 115 HM cycle genes were induced and 16 out of 115 genes were reduced, which corresponded to 10 upregulated and 10 down-regulated signal pathways, respectively (Figure 5A). We removed genes with contradictory expression pattern (found induced and reduced in the same disease category) from Figure 5 but discussed here. We identified 6 induced genes

(CHDH, KMT2A, KMT2B, PROM6, SETD1A and ICMT) and 4 reduced genes (BHMT,

KMT2D, METTL6 and HNMT) in T cell in HFH, 5 induced genes (PRMT2, EMG1,

TRMT2B, AS3MT and NNMT) and 1 reduced gene (EZH1) in omental adipose tissue in obesity, 5 induced genes (PRMT7, GNMT, PNMT, NT5C2 and NT5M) in T cell and 8 reduced genes (AHCY, MTR, PRMT5, PCMT1, DIMT1, TRMT5, PEMT, ADK) in pancreatic islets β cell in T2DM, 8 induced genes (EHMT2, EZH1, PRMT5, DIMT1,

TRMT1, TRMT61A, COMT and ADA) and 11 reduced genes (KMT2A, KMT2D,

PRDM9, PRMT2, TGS1, TRMT13, DNMT3B, DPH5, HNMT, SLC28A2 and ADK) in obesity-T2DM liver tissue. Associated with these induced/reduced genes, 10 signal pathways were found upregulated and 10 were downregulated.

In category B (6 metabolite treatments), we found 13 out of 115 induced HM cycle genes and 29 out of 115 reduced HM cycle genes, which corresponded to 11 upregulated and 10 down-regulated signal pathways, in which 2 pathways were overlapped.

We identified 10 induced genes (DOT1L, NSD1, SUV39H2, CARM1, PRMT7, COQ3,

DIMT1, TRMT13, METTL6, SLC28A2) and 12 reduced genes (CBS, CTH, SHMT2,

MTHFR, MAT1A, KMT2A, SETD1B, PRMT2, COMT, SLC28A3, NT5C, NT5M) in embryonic kidney cell under the treatment of 450mg/dl glucose, 3 induced genes

(ALDH7A1, DNMT3B, SLC29A1) in human CD34+ endothelial progenitor cell under

27 the treatment of 25mM glucose, 5 induced genes (CTH, PLD1, SETDB2, ALKBH8,

TRMT13) in human vascular smooth muscle cell under Ox-LDL treatment for 24 hours,

8 reduced genes (SHMT1, PLD1, SUV39H1, PRMT6, ALKBH8, HENMT1, MRM2,

HNMT) in human aortic endothelial cell under Ox-PAPC treatment, 2 reduced genes

(DOT1L, PRMT7) in aortic smooth muscle cell under 100umol/L Hcy treatment, 4 induced genes (PLD2, EMG1, SLC29A1, ADA) and 16 reduced genes (PON1, MTR,

CAMKMT, DOT1L, NSD1, PRDM2, GNMT, LRTOMT, HENMT1, TRMT61A,

METTL6, HNMT, INMT, SLC28A1, SLC29A2, NT5C1B) in vitamin B12 deficiency treated adipose tissue. Associated with these induced/reduced genes, 11 signal pathways were found upregulated and 10 were downregulated. These results suggested that extended HM cycle in embryonic kidney cells are highly susceptible to high glucose stimulation; and vitamin B12 deficiency in adipocytes downregulates so many more genes than upregulates genes in the extended HM cycle.

In category C (4 vascular diseases), we found 29 out of 115 induced HM cycle genes and 10 out of 115 reduced HM cycle genes, which corresponded to 10 up-regulated and 11 down-regulated signal pathways, in which 4 pathways were overlapped. We identified 3 reduced genes (EZH1, DNMT3L, SLC29A2) in human atherosclerotic foam cell, 12 reduced genes (AHCY, CBS, SHMT2, SETD3, PRMT5, ICMT, PCMT1, DIMT1,

TRMT5, HNMT, INMT, NNMT) in human aortic tissue in AOD, 34 induced genes

(AHCY, CTH, PON1, MTR, SHMT1, BHMT, PLD1, ALDH7A1, ASH1L, EHMT1,

PRDM6, PRDM7, SETD3, SETD7, GAMT, GNMT, ICMT, PCMT1, COQ3, TRMT5,

METTL1, METTL3, METTL6, DNMT3A, DNMT3B, INMT, PEMT, SLC28A1,

SLC28A3, SLC29A4, NT5C1A, NT5C2, NT5C3, NT5E) in T cell of CAD, 2 induced

(PLD1, RNMT) and 2 reduced genes (TRMT13, HNMT) in ischemic stroke’s peripheral blood mononuclear cell. Associated with these induced/reduced genes, 10 signal pathways were found upregulated and 11 signal pathways were downregulated. These results suggested that extended HM cycle has more downregulation of genes in aortic diseases but has more upregulation of genes in coronary artery disease.

In category D (4 aging diseases), we found 30 out of 115 induced HM cycle genes and 9 out of 115 reduced HM cycle genes, which corresponded to 10 up-regulated and 9 down-regulated signal pathways, in which 2 pathways were overlapped. We identified 16 induced genes (AHCY, CBS, CTH, PLD1, EZH1, EZH2, KMT2B, KMT2D,

SMYD2, SUV39H1, SUV39H2, TRMT1, METTL3, DNMT1, SLC29A1, SLC29A2) and 5 reduced genes (PRDM2, PRMT2, ALKBH8, HNMT, TPMT) in human primary skin fibroblasts in HGPS, 4 induced genes (MTHFR, KMT2B, SLC28A1, ADA) and 2 reduced genes (SETD7, HNMT) in substantia nigra of PD, 2 induced genes (RNMT,

HNMT) in human hippocampus of AD, 15 induced genes (DOT1L, EZH1, SETD2,

SETDB1, SETDB2, PRMT5, EMG1, RNMT, TRMT1, TRMT5, METTL1, METTL6,

NNMT, SLC28A3, NT5C) and 6 reduced genes (MTHFR, KMT2A, LRTOMT, INMT,

SLC29A1, NT5E) in human skeletal muscle in OSIMOF. These results suggested that extended HM cycle has extensive modulation of genes in three neurodegenerative diseases and one infectious disease. By comparison, HGPS and OSIMF have much more changes than that in PD and AD. Associated with these induced/reduced genes, 10 signal pathways were found upregulated and 9 signal pathways were downregulated.

In category E (5 autoimmune diseases), we found 23 out of 115 induced HM cycle genes and 18 out of 115 reduced HM cycle genes, which corresponded to 10 up-

29 regulated and 10 down-regulated pathways but no overlapped pathway. We identified 6 induced genes (AHCY, SHMT1, ALDH7A1, MRM2, DNMT3L, SLC29A1) and 8 reduced genes (SHMT2, KMT2D, PRMT7, DPH5, HNMT, PEMT, SLC29A2, SLC29A3) in human scalp in psoriasis, 18 induced genes and 7 reduced gene (CTH, ASH1L,

KMT2C, TRDMT1, HNMT, NNMT, ADK) in human mesenchymal stem cell in osteoarthritis, 3 induced genes (SETD3, FTSJ1, PNMT) in RA Jurkat T cell, 1 induced gene (CMTR1) and 10 reduced genes (SHMT1, MTHFR, KMT2B, KMT2D, PRDM2,

PRMT2, GNMT, ICMT, COMT, HNMT) in SLE monocyte, 1 induced gene (TPMT) in human distal lung fibroblast. These results suggestted that extended HM cycle has more induced genes in human osteoarthritis but has more reduced genes in SLE. Associated with these induced/reduced genes, 10 signal pathways were found upregulated and 10 were downregulated.

In category F (5 digestive cancers), we found 20 out of 115 induced HM cycle genes and 31 out of 115 reduced HM cycle genes, which corresponded to 10 up-regulated and 10 down-regulated pathways, in which 3 pathways were overlapped. We identified 9 induced genes (CBS, SHMT2, EHMT2, EZH2, DNMT3B, SLC28A3, SLC29A2, ADA) and 13 reduced genes (MTHFR, EZH1, KMT2A, PRDM6, SETD7, PRMT2, GNMT,

AS3MT, HNMT, INMT, NNMT, SLC29A1, NT5E) in esophageal tissue in ESCC, 3 induced genes (EHMT2, NNMT, SLC28A3) and 12 reduced genes (CTH, SHMT1,

MTHFR, BHMT, PRDM7, GAMT, METTL6, ASMT, SLC28A1, SLC28A2, SLC29A2,

ADA) in gastric tissue in GA, 9 induced genes (PLD2, PRDM2, PRDM6, SMYD3,

TRMT1, HNMT, SLC28A3, SLC29A1, NT5C2) and 18 reduced genes (AHCY, CTH,

SHMT1, PLD1, SUV39H1, SUV39H2, PRMT5, PRMT6, ICMT, PCMT1, COQ3,

TRMT61B, DNMT1, PEMT, TPMT, SLC28A1, SLC28A2, ADA) in pancreatic tissue in

PDA, 1 induced gene (RNMT) and 15 reduced genes (CBS, CTH, PON1, SHMT1,

BHMT, MAT1A, PLD1, CHDH, ALDH7A1, GAMT, GNMT, INMT, NNMT, PEMT,

SLC28A1) in liver tissue in IC, 16 induced genes (AHCY, SHMT2, ALDH7A1, EHMT2,

EZH2, NTMT1, DIMT1, HENMT1, TGS1, TRMT2B, TRMT5, TRMT61B, DNMT3A,

COMT, SLC29A1, ADK) in colon tissue in CC. These results suggested that extended

HM cycle has more reduced genes than induced genes in human digestive cancers except colon cancer. Associated with these induced/reduced genes, 10 signal pathways were found upregulated and 10 were downregulated.

In category G (6 other cancers), we found 27 out of 115 induced HM cycle genes and 26 out of 115 reduced HM cycle genes, which corresponded to 10 up-regulated and

11 down-regulated pathways, in which only 1 pathway was overlapped. We identified 16 induced genes (AHCY, CBS, SHMT2, CHDH, EZH2, SETDB1, SMYD2, ICMT,

TRMT1, TRMT61A, DNMT1, DNMT3B, SLC28A3, SLC29A1, SLC29A2, ADA) and

19 reduced genes (CTH, PON1, PLD1, ALDH7A1, DOT1L, EZH1, PRDM2, SETD7,

SETDB2, GAMT, GNMT, LRTOMT, ALKBH8, TRMT44, METTL6, AS3MT, DPH5,

HNMT, NT5E) in ovary tissue in OC, 6 induced genes (CHDH, EZH2, SMYD3, PRMT6,

FTSJ1, SLC29A2) and 1 reduced gene (SLC29A4) in prostate tissue in PC, 4 induced genes (CTH, PLD1, KMT2C, TRMT61A) and 4 reduced genes (PRMT2, DIMT1,

TRMT44, PEMT) in breast tissue in BA, 10 induced genes (AHCY, MARS, SHMT2,

PLD1, EZH2, COQ3, TGS1, DNMT3A, SLC29A4, NT5C3) and 5 reduced genes

(MTHFR, PRDM6, GNMT, INMT, PNMT) in lung tissue in NSCLC, 7 induced genes

(MAT1A, EZH2, NSD1, COQ3, DNMT3A, DNMT3B, ADK) and 4 reduced genes

(CTH, MTHFR, PRDM6, INMT) bladder tissue in BC, 5 induced genes (EZH2, SETD7,

SMYD2, NNMT, SLC29A4) and 9 reduced genes (AHCY, CTH, SHMT1, BHMT,

CHDH, ALDH7A1, GAMT, PEMT, TPMT) in kidney tissue in bladder tissue in CCRCC.

These results suggested that extended HM cycle has extensive modulation of genes in ovary carcinoma; in contrast, extended HM cycle has much less modulation of genes in prostate cancer. Associated with these induced/reduced genes, 10 signal pathways were found upregulated and 11 were downregulated.

Taken together, our analyses indicate that 35 pathological conditions modulate the expression changes of 115 extended HM cycle gene including 84 SAM/SAH-dependent

MT and result in 11 upregulated-only pathways, 3 downregulated-only pathways and 24 pathways with dual exchange pattern (Table 4). The extended HM cycle may sense the metabolic change and modulates as many as 38 signal pathways in response to 35 pathological conditions. The overlapped pathways suggested some different, perhaps opposite, molecular mechanisms are involved in this disease category. Our results have demonstrated that extended HM cycle is an extremely powerful metabolic sensor system in recognizing the concentration changes of different metabolic risk factors in as much as

35 disease conditions, mediating a very broad of methylation-regulated pathogenic signal.

This receptor-independent MADS recognition contributes to SAM/SAH-responsive signal pathways, which is much stronger than that of potential Hcy receptors such as N- methyl-D-aspartate receptor (NMDAr) in neuron system [72].

Figure 6A. Genes and pathways changed in 5 metabolic disease

Up-regulated pathway (10) Gene Changes Threshold -log(p-value) CHDH PNMT Adenosine Nucleotides Degradation II GNMT EMG1 Purine Nucleotides Degradation II (Aerobic) Guanosine Nucleotides Degradation III COMT TRMT1 Urate Biosynthesis/Inosine 5'-phosphate Degradation NNMT TRMT2B NAD Salvage Pathway II 19 genes Noradrenaline and Adrenaline Degradation EHMT2 TRMT61A induced Choline Degradation I KMT2B ASMT L-DOPA Degradation Catecholamine Biosynthesis PRDM6 NT5C2 Serotonin and Melatonin Biosynthesis SETD1A NT5M Ratio PRMT7 ADA ICMT Down-regulated pathway (10) Threshold -log(p-value) AHCY TGS1 Methionine Salvage II (Mammalian) BHMT TRMT13 Superpathway of Methionine Degradation DNMT3B TRMT5 Adenine and Adenosine Salvage VI Diphthamide Biosynthesis PEMT METTL6 16 genes Folate Transformations I MTR DPH5 reduced Glycine Betaine Degradation Histamine Degradation KMT2D HNMT Methionine Degradation I (to Homocysteine) PRDM9 SLC28A2 Cysteine Biosynthesis III (mammalia) DNA Methylation and Transcriptional Repression Signaling PCMT1 ADK Ratio

Figure 6B. Genes and pathways changed in 6 metabolite treatments

Up-regulated pathway (11) -log(p-value) Threshold Gene Changes Choline Degradation I Lysine Degradation II ALDH7A1 DIMT1 Lysine Degradation V DNMT3B EMG1 Adenine and Adenosine Salvage III PLD2 TRMT13 Purine Ribonucleotides Degradation to Ribose-1-phosphate 13 genes Choline Biosynthesis III SETDB2 SLC28A2 induced Adenosine Nucleotides Degradation II SUV39H2 SLC29A1 Histamine Degradation Purine Nucleotides Degradation II (Aerobic) CARM1 ADA Ubiquinol-10 Biosynthesis (Eukaryotic) COQ3 Oxidative Ethanol Degradation III CBS SUV39H1 Ratio PON1 PRMT2 MAT1A PRMT6 GNMT LRTOMT COMT HENMT1 Down-regulated pathway (10) INMT MRM2 Threshold -log(p-value) MTR TRMT61A Folate Transformations I 29 genes Superpathway of Methionine Degradation SHMT1 HNMT reduced Cysteine Biosynthesis III (mammalia) SHMT2 SLC28A1 Guanosine Nucleotides Degradation III MTHFR SLC28A3 Urate Biosynthesis/Inosine 5'-phosphate Degradation Adenosine Nucleotides Degradation II CAMKMT SLC29A2 L-DOPA Degradation EHMT1 NT5C1B Glycine Biosynthesis I Purine Nucleotides Degradation II (Aerobic) KMT2A NT5C Methionine Degradation I (to Homocysteine) PRDM2 NT5M Ratio SETD1B

Figure 6C. Genes and pathways changed in 4 vascular diseases

Up-regulated pathway (10) Gene Regulation Threshold CTH SETD7 -log(p-value) Guanosine Nucleotides Degradation III PON1 COQ3 Urate Biosynthesis/Inosine 5'-phosphate Degradation BHMT RNMT Adenosine Nucleotides Degradation II ALDH7A1 DNMT3A Purine Nucleotides Degradation II (Aerobic) NAD Salvage Pathway II GAMT METTL1 Superpathway of Methionine Degradation GNMT METTL3 Methionine Salvage II (mammalia) DNMT3B METTL6 Folate Transformations I 29 genes Glycine Betaine Degradation PEMT SLC28A1 induced Cysteine Biosynthesis III (mammalia) MTR SLC28A3 SHMT1 SLC29A4 Ratio PLD1 NT5C1A Down-regulated pathway (11) -log(p-value) ASH1L NT5C2 Cysteine Biosynthesis III (mammalia) EHMT1 NT5C3 Superpathway of Methionine Degradation PRDM6 NT5E Cysteine Biosynthesis/Homocysteine Degradation PRDM7 Glycine Biosynthesis I dTMP De Novo Biosynthesis CBS DIMT1 Folate Polyglutamylation NNMT TRMT13 Superpathway of Serine and Glycine Biosynthesis I 10 genes SHMT2 DNMT3L Folate Transformations I reduced Glycine Betaine Degradation EZH1 HNMT Histamine Degradation PRMT5 SLC29A2 Methionine Degradation I (to Homocysteine) Ratio

Figure 6D. Genes and pathways changed in 4 aging diseases

Gene Changes Up-regulated pathway (10) Threshold -log(p-value) AHCY SUV39H2 Cysteine Biosynthesis III (mammalia) CBS PRMT5 Superpathway of Methionine Degradation Cysteine Biosynthesis/Homocysteine Degradation CTH EMG1 Methionine Degradation I (to Homocysteine) NNMT RNMT Adenosine Nucleotides Degradation II Purine Nucleotides Degradation II (Aerobic) PLD1 TRMT1 Adipogenesis pathway L-cysteine Degradation II DOT1L TRMT5 Adenine and Adenosine Salvage III EZH1 DNMT1 Purine Ribonucleotides Degradation to Ribose-1-phosphate EZH2 METTL1 30 genes induced Ratio KMT2B METTL3 KMT2D METTL6 SETD2 SLC28A1 Down-regulated pathway (9) SETDB1 SLC28A3 -log(p-value) Threshold SETDB2 SLC29A2 L-DOPA Degradation SMYD2 NT5C Guanosine Nucleotides Degradation III Urate Biosynthesis/Inosine 5'-phosphate Degradation SUV39H1 ADA Adenosine Nucleotides Degradation II INMT LRTOMT Purine Nucleotides Degradation II (Aerobic) NAD Salvage Pathway II KMT2A ALKBH8 Dopamine Degradation PRDM2 TPMT 9 genes Noradrenaline and Adrenaline Degradation reduced Nicotine Degradation II SETD7 NT5E Ratio PRMT2

Figure 6E. Genes and pathways changed in 5 autoimmune diseases

Up-regulated pathway (10) -log(p-value) Gene Changes Threshold AHCY PRMT1 Methionine Degradation I (to Homocysteine) CHDH NTMT1 Cysteine Biosynthesis III (mammalia) ALDH7A1 PNMT Superpathway of Methionine Degradation Choline Degradation I GAMT CMTR1 Noradrenaline and Adrenaline Degradation DOT1L FTSJ1 Glycine Degradation (Creatine Biosynthesis) EHMT2 MRM2 23 genes Catecholamine Biosynthesis Lysine Degradation II KMT2A DNMT3L Induced Lysine Degradation V NSD1 TPMT Guanosine Nucleotides Degradation III SETD1A SLC28A1 Ratio SETD3 SLC29A1 SMYD2 NT5E Down-regulated pathway (10) -log(p-value) SMYD3 Threshold CTH PRDM2 Folate Transformations I L-cysteine Degradation II GNMT PRMT2 Adenine and Adenosine Salvage VI COMT PRMT7 L-DOPA Degradation NNMT ICMT Cysteine Biosynthesis/Homocysteine Degradation 18 genes PEMT TRDMT1 Glycine Biosynthesis I reduced Diphthamide Biosynthesis SHMT2 DPH5 dTMP De Novo Biosynthesis MTHFR HNMT Folate Polyglutamylation ASH1L SLC29A3 Superpathway of Serine and Glycine Biosynthesis I KMT2C ADK Ratio

Figure 6F. Genes and pathways changed in 5 digestive cancers

Gene Changes Up-regulated pathway (10) DNMT3B HENMT1 -log(-log(p-pvalue)-value) COMT RNMT Threshold SHMT2 TGS1 DNA Methylation and Transcriptional Repression Signaling PLD2 TRMT1 Adenine and Adenosine Salvage VI L-DOPA Degradation EHMT2 TRMT2B 20 genes Glycine Biosynthesis I EZH2 TRMT5 Induced dTMP De Novo Biosynthesis PRDM2 DNMT3A Folate Polyglutamylation Superpathway of Serine and Glycine Biosynthesis I SMYD3 SLC28A3 Folate Transformations I NTMT1 NT5C2 Glycine Betaine Degradation DIMT1 ADK Guanosine Nucleotides Degradation III CTH SUV39H2 Ratio PON1 PRMT2 BHMT PRMT5 MAT1A PRMT6 Down-regulated pathway (10) -log(p-value) CHDH AS3MT Threshold -log(p-value) GAMT ICMT Superpathway of Methionine Degradation GNMT PCMT1 Cysteine Biosynthesis III (mammalia) Methionine Degradation I (to Homocysteine) INMT COQ3 31 genes Folate Transformations I PEMT DNMT1 reduced Glycine Betaine Degradation SHMT1 METTL6 L-cysteine Degradation II MTHFR ASMT Choline Degradation I Glycine Degradation (Creatine Biosynthesis) EZH1 TPMT Cysteine Biosynthesis/Homocysteine Degradation KMT2A SLC28A1 Glycine Biosynthesis I PRDM7 SLC28A2 Ratio SETD7 NT5E Ratio SUV39H1

Figure 6G. Genes and pathways changed in 6 other cancers

Gene Changes CBS COQ3 Up-regulated pathway (10) MAT1A FTSJ1 Threshold -log(p-value) DNMT3B TGS1 Cysteine Biosynthesis III (mammalia) NNMT TRDMT1 DNA Methylation and Transcriptional Repression Signaling MARS TRMT61A Superpathway of Methionine Degradation SHMT2 DNMT1 Adenosine Nucleotides Degradation II EZH2 DNMT3A 27 genes Purine Nucleotides Degradation II (Aerobic) KMT2C SLC28A3 induced Methionine Degradation I (to Homocysteine) NSD1 SLC29A1 Adenine and Adenosine Salvage VI SETDB1 SLC29A2 Cysteine Biosynthesis/Homocysteine Degradation SMYD2 NT5C3 Glycine Biosynthesis I SMYD3 ADA S-adenosyl-L-methionine Biosynthesis PRMT6 ADK Ratio ICMT PON1 PRDM6 Down-regulated pathway (11) -log(p-value) BHMT SETDB2 ALDH7A1 PRMT2 Ratio Noradrenaline and Adrenaline Degradation GAMT AS3MT Folate Transformations I GNMT LRTOMT Glycine Betaine Degradation TRMT44 PNMT 26 genes Histamine Degradation INMT ALKBH8 Dopamine Degradation reduced PEMT DIMT1 Choline Degradation I SHMT1 METTL6 L-DOPA Degradation Glycine Degradation (Creatine Biosynthesis) MTHFR DPH5 Glycine Biosynthesis I DOT1L HNMT Diphthamide Biosynthesis EZH1 TPMT Methionine Salvage II (mammalia) PRDM2 NT5E Ratio Figure 6. Genes and pathways changed in 7 disease condition categories (A to G).

We examined gene expression of 115 enzymes in extended HM cycle in microarray datasets of 35 human disease conditions in NIH-NCBI GEO database

(https://www.ncbi.nlm.nih.gov/geo/). These 35 disease conditions were classified in 7 categories (A to G). Gene expression changes are listed in the tables. Signal pathways regulated by gene expression changes were analyzed using Ingenuity Pathway Analysis

(IPA, http://www.ingenuity.com/) Core Analysis. Up- and downregulated pathways are described as listed. Bar graph describes statistic power of identified pathway and expressed as -log (p-value). Threshold in the graph describes minimum significance levels (p=0.05). Leaner curve in the graph describes the ratio of the number of identified genes divided by the total number of genes involved in this pathway from the IPA knowledgebase. Abbreviations for gene and enzyme names are explained in Table 1.

Table 4. Classification of SAM/SAH-responsive signal pathways in human disease category

Disease category With down- # Pathways With up-regulated regulated pathways pathways 1 Serotonin and Melatonin Biosynthesis Metabolic Disease (Dis) 2 Adipogenesis pathway Aging Dis 3 SAM Biosynthesis Other Cancer (Can) Metabolite Treatment 4 Oxidative Ethanol Degradation III (Treat) Ubiquinol-10 Biosynthesis 5 Metabolite Treat (Eukaryotic) 6 Choline Biosynthesis III Metabolite Treat Metabolic, Autoimmune 7 Catecholamine Biosynthesis Dis Aging Dis, Metabolite 8 Adenine and Ade Salvage III Treat Purine Ribonucleotides Degradation to Aging Dis, Metabolite 9 Ribose-1-phosphate Treat Autoimmune Dis, 10 Lys Degradation I Metabolite Treat Autoimmune Dis, 11 Lys Degradation V Metabolite Treat 12 Nicotine Degradation II Aging Dis Aging Dis, 13 Dopamine Degradation Other Can Metabolic, 14 Diphthamide Biosynthesis Autoimmune Dis, Other Can

Table 4. Summary of HM cycle gene expression-modulated pathways in human disease category. We summarized signal pathways that are modulated by altered HM cycle gene expression changes in 7 categories. We found that 11 pathways are only upregulated, and 3 pathways are only down-regulated in these diseases.

3.4 Thirty-three HM cycle-regulated and SAM/SAH-dependent H3/H4 histone methylations are changed in human disease

HHcy, vitamin B12 deficiency, SLE, psoriasis, atherosclerosis, aortic occlusive disease, T2DM, and Ox-PAPC may induce site-specific histone hypomethylations by downregulating specific SAM/SAH-dependent histone methyltransferases — Based on the identified histone MT gene expression changes and their corresponding modification sites (Table 2), we defined H3/H4 histone methylation pattern (Figure 7A and 12B) in

29 human disease conditions which are presented by code (Figure 7C). Percentage indicates the frequency of each modification in total methylation changes modified.

Codes for disease with individual hypermethylation changes are placed above the histone bar, whereas for that with individual hypomethylation changes below the histone bar.

In Table 2, 6 histone arginine MT methylate four arginine residues (Rs) such as

R2 (number 2 amino acid residue from the N-terminus), R8, R17, and R26 on the N- terminus of histone 3 and one arginine residue R3 on the N-terminus of histone 4.

Furthermore, 30 histone lysine MT methylate six lysine residues (Ks) including 5 lysine residues K4, K9, K27, K36, K79 on the N-terminus of histone 3 and one lysine residue

K20 on the N-terminus of histone 4. We have analyzed the expression changes of all the

36 histone MT in 35 pathological conditions and found 5 out of 35 pathological conditions such as IC, AD, asthma, IS and high glucose (25 mM) treatment induce no expression changes in histone MT; Parkinson’s disease induces KMT2B and reduced

SETD7, but these two enzymes both modified H3K4me1. The rest 29 disease conditions cause SAM/SAH-dependent H3 and H4 histone methylation changes (Figure 7A and

7B). Also, the rest 29 disease conditions changed 34 histone MT (Figure 8). These histone modifications are termed as HM cycle-regulated and SAM/SAH-dependent methylation.

We recently reported that metabolic diseases downregulate the majority of histone modification enzymes, making a few upregulated enzymes novel therapeutic targets [29]; and that proatherogenic lipid lysophosphatidylcholine (LysoPC) induce human aortic endothelial cell activation via increasing site-specific histone 3 lysine 14 acetylation [57], which is mutually exclusive from histone methylation at lysine and arginine residues in the N-terminus of histones [29]. We also reported that Hcy inhibits cyclin A transcription and cell growth by inhibiting DNA methylation through suppression of DNMT1 in endothelial cells [23], suggesting that DNA hypomethylation is a key biochemical mechanism responsible for Hcy-promoted cardiovascular disease [73]. Thus, we hypothesize that some diseases including HHcy may modulate site-specific histone methylations via the actions of SAM/SAH-dependent MT.

We mapped all the potential histone methylation changes induced by 29 pathological conditions. As shown in Figure 7A, we found that first, pathological conditions increase 84.5% of methylations on histone 3 residues but only increase 15.5% of methylations on histone 4 residues; second, pathological conditions decrease 79.5% of methylations on histone 3 residues but only increase 20.5% of methylations on histone 4 residues; third, most of methylation increases in pathological conditions focus (>10%) on

H3K4 (15.5%), H3K9 (17.2%), H3K27 (13.8%), H3K36 (15.5%) and H4R3 (10.3%); fourth, most of methylation decreases in pathological conditions focus (>10%) on H3R2

(13.6%), H3K4 (20.5%), H3R8 (18.2%), H3K9 (13.6%) and H4R3 (13.6%); fifth, 8

39 types of pathological conditions including HHcy, vitamin B12 deficiency, SLE, psoriasis, atherosclerosis, AOD, T2DM, and Ox-PAPC treatment may induce site-specific histone hypomethylations by reducing specific SAM/SAH-dependent histone MT gene expression level; while 9 pathological conditions (osteoarthritis, HFH, CAD, PC,

CCRCC, Ox-LDL treatment, CC and RA) may only induce site-specific histone hypermethylation. Interestingly, HHcy (#19 disease condition) decreases histone methylation on only three residues, H3R2, H3K9 and H4R3; atherosclerosis (#25 disease condition) decreases histone methylation on only one residue, H3K27; and T2DM β cells

(#29 disease condition) decreases histone methylation on only three residues, H3R2,

H3R8 and H4R3. These results are well correlated with our previous reports that HHcy has synergistic effects with hyperlipidemia [74] and T2DM [8, 75] and promotes inflammatory monocyte differentiation, vascular inflammation and atherosclerosis [6].

Our results here have demonstrated that metabolic disease-induced SAM/SAH-dependent hypomethylation mechanisms not only apply to DNA hypomethylation [23] and RNAs

[76] as we and others reported, respectively, but also apply to histone hypomethylation at site-specific manners similar to that we reported for H3K14 acetylation in human aortic endothelial cell activation induced by proatherogenic lipid lysoPC [57].

A. H3 methylation changes

Disease condition

n o

i Total %

a l

y % 8.6 15.5 5.2 17.2 1.7 1.7 13.8 15.5 5.2 84.5

h t

e 1, 2, 3 1, 2, 4, 9 1, 5, 6 1, 2, 4, 6 m

r 5, 9, 10 4, 5, 6 10, 13, 14 10, 14, 21 9, 10, 15 e

p Code 7, 8, 26 16, 17, 21 23, 24 24, 26

y 14, 28 10, 14, 22 9 9 2, 9, 10 H …

Histone 3 R2 … K4 … R8 K9 … R17 … R26 K27 … K36 K79

i t

a Code:

l 7, 11, 18 9, 10, 11 1,7,8, 9 7, 12, 15 22, 25 20, 27 15, 20

y h

t 19, 27, 29 12, 13, 14 12, 13 18, 19, 20

e 15, 16, 27 27, 29

m o

p %: 13.6 20.5 18.2 13.6 4.5 4.5 4.5 79.5

y H

B. H4 methylation changes C. Disease conditions: Code Name Code Name

Disease condition 1 Hutchinson-Gilford progeria syndrome 16 Gastric adenocarcinoma n

o 2 Osteoarthritis 17 Ox-LDL 24 hours i

t Total % a l %: 3 Heterozygote Family Hypercholesterolemia 18 Ox-PAPC

y 10.3 5.2 15.5 h

t 4 Coronary artery disease 19 Hcy (100umol/L) e

m 5 20

r 2, 5, 9, Prostate cancer Vitamin B12 deficiency e

p Code: 10, 14, 28 3, 4, 7 6 Clear cell renal cell carcinoma 21 Colon cancer y

H 7 Pancreatic ductal adenocarcinoma 22 Obesity … Histone 4 R3 K20 8 Breast adenocarcinoma 23 Non-small cell lung cancer

n 9 High glucose (450mg/dl) 24 Bladder cancer

i t

a Code: 7, 11, 18 13, 23, 24 10 25

l Old sepsis induced multiple organ failure Atherosclerosis y

h 19, 27, 29

t 11 Psoriasis 26 Rheumatoid arthritis e

m 12 Systemic lupus erythematosus 27 Aortic occlusive disease o %: 13.6 6.8 p 20.5 13 28

y Esophageal squamous cell carcinomas T2DM (Blood) H 14 Obesity-T2DM 29 T2DM (β cell) 15 Ovarian carcinoma

Figure 7. HM cycle-regulated and SAM/SAH-dependent H3/H4 histone methylations are changed in human disease. We examined gene expression of 115 enzymes in extended HM cycle in microarray datasets of 35 human disease conditions in NIH-NCBI GEO database (https://www.ncbi.nlm.nih.gov/geo/). Based on the identified expression changes of HM cycle gene and histone MT gene and their corresponding modification site (Table 2), we defined induced MT gene as hypermethylation and reduced MT gene as hypomethylation on their corresponding histone methylation site. We then determined H3/H4 histone methylation pattern (A) in 29 human disease conditions (C). A. H3 methylation changes. H3 histone methylation are changed at H3- R2, K4, R8, K9, R17, R26, K27, K36, K79 in human disease. Disease condition are presented by code. Percentage indicates the frequency of each modification in total methylation changes modified. Codes for disease with individual hypermethylation

41 changes are placed above the histone bar, whereas for that with individual hypomethylation changes below the histone bar. B. H4 methylation changes in human diseases. H4 histone methylation are changed at H4-R3 and K20 in human disease. C. Disease conditions

Figure 8. Histone MT regulation in 29 disease conditions

)

) L

Disease s

/ o

o C

conditions C

F i

i r

L y

t T β

C e

T - -

D A m u

H D

(29) E

r f

P -

C A

A A

C A

C C

C c

0 0

F A

M r

C M M

a R e

B t P

P O G C B R

5 0

o S - d

x b

H C P

D D

4 1

s E

Histone H

2 2

( (

O O N

t 2

O P

O T T

G y b

MT h

A V (36) H Histone-lysine MT (30): ASH1L - + CAMKMT - DOT1L + + + - - EHMT1 + + - EHMT2 + + + + + + EZH1 + - + - - - EZH2 + + + + + + + KMT2A + + - - - - KMT2B + + + - KMT2C - + KMT2D + + - - - - KMT2E KMT5A NSD1 + + - + PRDM2 - + - - - PRDM6 + + + - - - PRDM7 + - PRDM9 - SETD1A + + SETD1B - SETD2 + SETD3 + + + - SETD7 + + - - SETDB1 + + SETDB2 + - + SETMAR SMYD2 + + + + SMYD3 + + + SUV39H1 + - - SUV39H2 + - + Histone-arginine MT (6): CARM1 + PRMT1 + PRMT2 ------PRMT5 - + + - - PRMT6 + - - PRMT7 + - - +

Figure 8. Histone MT regulation in 29 disease conditions. We examined gene expression of 36 histone methyltransferase in 29 human disease conditions as discussed in Figure 7.

We found the 33 out of 36 Histone MT gene expression level are modulated in 29 disease conditions. + indicated that the gene was induced and – indicated that the gene is reduced.

Bolded letter highlighted genes with changed expression.

CHAPTER 4 RESULTS II

4.1 RNA-Seq analysis for mouse CD40+ and CD40- monocyte (MC)

CD40 is a cell surface molecule which is expressed on antigen presenting cells

(APC) such as monocyte, macrophage, dendritic cells and neutrophils. According to our previous study, CD40+ MC is a stronger pro-inflammatory MC subset compared to intermediated MC and a reliable biomarker for CKD severity [9, 46]. In order to establish gene expression profiles in different monocyte subsets and CD40 function on methylation regulated pathogenic signaling, Monocyte isolation (Figure 9) and RNA-seq analysis are performed for CD40+ monocytes from mouse peripheral blood.

After sorting out the CD40+ and CD40- MC, RNAs are isolated from 2 different types of MCs and RNA sequencing analysis was performed. We have 2 samples in each group as technically duplicates. CD40+ and CD40- MC are successfully separated as shown in Figure 10.

A peripheral blood WT mouse Collection FACS Sorting RNA isolation RNA sequencing 0 12 wks

B WBC Live cells

88.4% 86.8% Live cells Singlets

Singlets Monocytes CD40+ 8.2%

MC 28.6%

CD40- 34.3%

Figure 9. Sorting strategy for CD40+ and CD40- MC. Peripheral blood are collected from 12 weeks wide type mice. By using gating strategy in flow cytometry, CD40+ and

CD40- MC are isolated for further RNA-sequencing analysis. A indicated our flow cytometry strategy flow. B indicated the gating strategy.

4.2 Differentially expressed genes in CD40+ vs CD40- MC

After the RNA-sequence analysis, we totally got 24,341 annotated genes. By using the cut off strategy of FDR less than 0.05, we got 2,093 genes with significant changes. Those 2,093 significantly changed genes could be divided into 15 functional groups by Ingenuity Pathway Analysis software, including enzyme (357 genes), transcription regulator (200 genes), kinase (142 genes), transporter (119 genes), transmembrane receptor (115 genes), peptidase (62 genes), phosphatase (49 genes), G- protein coupled receptor (45 genes), ion channel (23), translation regulator (17 genes), cytokine (10 genes), growth factor (8 genes), ligand-dependent nuclear receptor (6 genes) microRNA (1 gene) and other ungrouped genes (938 genes). Those information are summarized in Table 5, Table 6 and Figure 11.

After digging out 2,092 significantly changed genes in CD40+ MC, we further studied genes with 2 folds change. There are 1,093 genes we finally dig out using a cut off strategy by FDR less than 0.05 and absolute fold change over 2. Volcano plot (Figure

12) showing log2(Fold change, FC) and -log10 (P value) of 24,341 genes. Significantly changed genes are indicated by red dots.

In Figure 13, the heat map showed that among all the 1,093 genes, 589 genes are upregulated in CD40+ MC while 504 genes are downregulated in CD40- MC.

CD40+ MC CD40- MC

Figure 10. PCA plot results. CD40+ and CD40- MC were successfully separated in PCA plot. Red dots indicated the CD40+ MC and the blue dots indicated the CD40-. These two groups of MC are well separated.

Table 5. Summary of 24,341 annotated genes

Summary Gene # Annotated gene 24,341 Genes with FDR ≤ 0.05 2,092 Upregulated 1008 Downregulated 1084 Genes with Fold Change ≥ 2 1,093 upregulated 589 downregulated 504

Table 5. Summary of 24,341 annotated genes. This table summarized the all annotated genes from our RNA seq results. Among 24,341 annotated genes, 2,092 genes are significantly changed in CD40+ MC.

Table 6. 15 functional groups of DEGs

Group # Group name Gene #

1 microRNA 1 2 ligand-dependent nuclear receptor 6 3 growth factor 8 4 cytokine 10 5 translation regulator 17 6 ion channel 23 7 G-protein coupled receptor 45 8 phosphatase 49 9 peptidase 62 10 transmembrane receptor 115 11 transporter 119 12 kinase 142 13 transcription regulator 200 14 enzyme 357 15 other 938

Table 6. 15 functional groups of DEGs. 2,092 differentially expressed genes with the

FDR less than 0.05 can be divided into 15 functional groups based on the IPA software.

Figure 11. Differentially regulated 15 gene functional groups in CD40+ MC. 15 differentially regulated gene functional groups are shown by pie chart and bar graph.

)

( FC ≥ 2

0 FC <2

l -

log2FC

Figure 12. Volcano plot showing log2(Fold change, FC) and –log10 (P value) of

24,341 genes. Red dots indicate significantly changed (FDR ≤ 0.05) genes by more than

2-folds (1093 genes) in CD40+ monocytes.

CD40+ CD40-

1 2 1 2 _ _ _ _ s s g g o o e e p p n n 0 0 0 0 4 4 4 4 D D D D C C C C

Figure 13. Heat map result. Genes that are significantly changed by more than 2 folds

(1093 genes, CD40+ MC vs CD40- MC).

4.3 Differentially expressed genes in HM cycle and trained immunity pathways

According to our previous study, DNA hypomethylation on CD40 promoter and suppression of DNA MT DNMT1 contributed to Hcy–induced inflammatory CD40+ MC differentiation in CKD, so we considered CD40 as a prototypic homocysteine-methionine cycle regulated master gene and HHcy induced hypomethylation target gene. In HM cycle, CD40 is a target molecule to induce the most of the Hcy metabolic enzymes as well as MT, which can modulate the methylation-regulated pathological signaling. By analyzing the 115 HM cycle gene expression level in CD40+ MC, we found CD40+ monocyte can induce the majority of MT in the cycle, which could regulate the Hcy metabolism and Hcy-related methylation. However, DNA MTs had no significant change, which indicates that CD40+ MC may induce DNA hypomethylation caused by the elevation of HHcy. The significantly changed genes in HM cycle are listed in Table 7.

Trained immunity is a newly proposed concept which indicates the immune memory of the innate immune system and involves the epigenetic programming, resulting in changes in their metabolic and phenotypical behavior that enable a stronger immune response to secondary stimuli. Since monocyte is the innate immune cell, we come up with a hypothesis that if CD40+ MC could induce the trained immunity reaction and cause a stronger immune response. We analysis the genes changed in three major pathways of trained immunity including glycolysis pathway, Acetyl-CoA generation pathway and mevalonate pathway and found that CD40+ MC can induce the trained immunity especially from Acetyl-CoA generation pathway and mevalonate pathway

(Table 8). Since Hcy is a metabolite and an independent risk factor for many diseases, this finding gave us a clue that Hcy induced hypomethylation and inflammatory MC

51 differentiation may through the trained immunity pathway and finally causes many pathophysiological results (Table 8).

Table 7. CD40+ MC gene changes in HM cycle

Gene symbol Fold Change P value Homocysteine metabolism Shmt2 2.28154183 0.00025 Pld1 -2.120291 0.00035 Histone-lysine MT Kmt2a 1.28104312 0.04735 Setd1b -1.4862232 0.0012 Smyd2 5.97475637 0.0114 Histone-arginine MT Prmt2 2.21302687 0.02745 Prmt5 1.59841213 0.03345 Prmt7 2.2103426 0.00335 RNA MT Alkbh8 2.00439901 0.0105 Cmtr1 1.51534945 0.0424 Ftsj3 1.45365467 0.01555 Trmt1 1.69123796 0.00405 Trmt61a 2.35235691 0.01305 Mettl1 4.09353103 0.0169 Ungrouped MT Dph5 2.22736476 0.0009 DNA MT Dnmt1 -1.1143454 0.3987 Dnmt3a -1.0117899 0.9323 Dnmt3b -1.680393 1 Dnmt3l 5.08536832 1 Adenosine metabolism Slc28a2 2.20240098 0.00005 Slc29a1 -1.6883239 0.00995 Slc29a3 1.47728969 0.0109 Nt5e 2.55560531 0.00225 Ada 3.47249977 0.00465

Table 7. CD40+ MC gene changes in HM cycle. CD40+ MC upregulated the majority of MT, but not induced DNA MT.

Table 8. CD40+ MC gene changes in trained immunity pathways

Trained immunity pathway Gene symbol Fold Change p value ALDOC 3.43 0.020 BPGM 2.41933404 0.000 ALDH2 1.93012932 0.00005 ACSS1 1.79948368 0.00005 PDHB 1.75287636 0.0145 ADPGK 1.50862674 0.0215 ADH5 1.48325364 0.03335 ALDH3A2 -1.3457364 0.04425 Glycolysis Pathway PDHA1 -1.3507736 0.0385 HK2 -1.6940122 0.00025 PFKFB4 -1.8548133 0.00005 PFKP -1.9886381 0.00005 ACSS2 -2.2516083 0.00585 LDHB -3.0409573 0.01245 HK3 -3.0933317 0.00005 ENO3 -3.46073 0.00005 GOT1 2.14413523 0.01655 ALDH2 1.93012932 0.00005 ACSS1 1.79948368 0.00005 Acetyl-CoA generation enzyme FASN 1.57756456 0.00125 ACAT1 1.45946725 0.0249 ACSS2 -2.2516083 0.00585 ACE -4.2005396 0.00005 ACAT1 1.45946725 0.0249 Mevalonate pathway IDI1 2.20897539 0.03515

Table 8. CD40+ MC gene changes in trained immunity pathways. CD40+ monocytes activated trained immunity pathways especially in Acetyl-CoA generation and mevalonate pathway.

CHAPTER 5 DISCUSSION

Regardless the significant progress has been made in the field, we identified a few key questions to address here: First, we asked whether in addition to enzymatic generation of SAH, Hcy, methionine, and SAM, the extended HM cycle could serves as a new sensor for SAH/Hcy increase and SAM/SAH ratio changes; second, since that SAH is generated after SAM donates methyl group for intracellular methylation reactions including DNA methylation, RNA methylation and protein/histone methylation; and that

SAH serves as the inhibitor for SAM-mediated global methylation, then all the methyltransferases are the parts of extended methionine cycle. How all the enzymes in extended methionine cycle are regulated in pathological conditions. To address these important questions, we took a panoramic view at the expressions of 115 enzymes in extended HM cycle in 35 disease conditions including metabolic diseases, vascular diseases, aging diseases, autoimmune diseases and cancers. We have made the following important findings: 1) Extended HM cycle serves as a new sensor for hyperhomocysteinemia, methionine, S-adenosylmethionine, and S-adenosylhomocysteine;

2) A majority of HM cycle enzymes are located in cytosol, but 61 out of 84 SAM- dependent methyltransferases are located in nucleus; 3) The extended HM cycle may sense the metabolite changes and modulates as many as 38 signaling pathways in response to 35 pathological conditions; 4) HHcy, vitamin B12 deficiency, lupus, psoriasis, atherosclerosis, aortic occlusive disease, type II diabetes, and ox-PAPC may induce site- specific histone hypomethylations by downregulating specific SAM-/SAH dependent histone MT.

For the analysis, we used an experimental database mining approach that was pioneered and developed in our laboratory throughout the years [77-79]. By analyzing high precision gene expression data deposited in NIH-GEO Datasets we could determine the expression changes of HM cycle enzymes and 84 SAM-dependent methyltransferases under pathological conditions. Of course, the future works are needed to verify these findings with additional experiments.

HHcy and hypermethioninemia (HMet) are often connected genetically and non- genetically. Several genetic conditions have been identified including [80]: 1) HHcy due to cystathionine β-synthase (CBS) deficiency. At least 150 different mutations in the CBS gene have been characterized since 1964 when this deficiency was established. 2)

Deficient activity of methionine adenosyltransferases I and III (MAT I/III), the isoenzymes the catalytic subunit of which are encoded by MAT1A. Methionine accumulates because its conversion to SAM is impaired; 3) SAH hydrolase

(AHCY/SAHH) deficiency. Not being catabolized normally, SAH accumulates and inhibits many SAM-dependent methyltransferases, producing accumulation of SAM.

Non-genetic conditions have also been identified including: first, liver disease, which may cause HMet, mild, or severe; and second, low-birth-weight and/or prematurity, which may cause transient HMet [80]. By comparison, our report here has demonstrated that as many as 35 diseases and disease risk factor stimulations, the majority of which are not traditionally Hcy-related diseases, modulate extensively the expressions of HM cycle enzymes and 84 SAM-dependent methyltransferases, which leads to regulation of a broad spectrum of 38 different signaling pathways.

Those major findings presented here significantly improve our understanding on various diseases modulate methylation-based signaling in a very broad spectrum, which is complimentary to Hcy-, Met-, and Cys-related oxidative mechanisms. In summary, we would like to propose a new working model to integrate our new findings: first, risk factors associated with various diseases including metabolic diseases, neurodegenerative diseases, autoimmune diseases, and cancers act on DAMP receptors, MADS receptors, conditional DAMP receptors modulate the gene transcription of genes encoding HMC enzymes and all the 84 SAM-dependent methyltransferases; second, HM cycle with all the 84 SAM-dependent methyltransferases serves as a novel metabolic sensor for detecting concentration changes of Hcy, methionine, SAM and SAH and initiate a broad spectrum of 38 signaling pathways in modulating the pathogenesis of various diseases and physiological conditions; third, putative cytosol-nucleus SAM/SAH transfer are new therapeutic targets; and third, HHcy, vitamin B12 deficiency, lupus, psoriasis, atherosclerosis, aortic occlusive disease, type II diabetes, and ox-PAPC may induce site- specific histone hypomethylations by downregulating specific SAM-dependent histone methyltransferases. Our findings not only provide a novel insight on HHcy and other amino acid-related pathologies and new therapeutic targets but also make HM cycles as new methylation signaling hub for broad-spectrum of physiological and pathological conditions.

CD40 is a cell surface molecule which is expressed on antigen presenting cells such as monocyte, macrophage, dendritic cells and neutrophils. The costimulatory pair,

CD40 and CD40L, enhances T cell activation and induce chronic inflammatory disease.

Also, DNA hypomethylation on CD40 promotor induces inflammatory monocyte

56 differentiation in CKD. In our RNA seq analysis results, genes in CD40+ MC were differentially expressed in as many as 15 functional groups. Interestingly, CD40+ MC induced most of the HM cycle MT except for the DNA MT, which matches our previous result that CD40 induces DNA hypomethylation. Also, CD40 could cause trained immunity pathway change leading us to have further study on the metabolic pathway in trained immunity. In conclusion, we proposed CD40 as a prototypic HM cycle regulated master gene.

REFERENCES CITED

1. Kumar, A., et al., The metabolism and significance of homocysteine in nutrition

and health. Nutrition & metabolism, 2017. 14: p. 78.

2. Lee, M.E. and H. Wang, Homocysteine and hypomethylation. A novel link to

vascular disease. Trends in cardiovascular medicine, 1999. 9(1-2): p. 49-54.

3. Jensen, M.K., et al., Novel metabolic biomarkers of cardiovascular disease.

Nature reviews Endocrinology, 2014. 10(11): p. 659-72.

4. Lee, H.O., et al., S-adenosylhomocysteine hydrolase over-expression does not

alter S-adenosylmethionine or S-adenosylhomocysteine levels in CBS deficient

mice. Mol Genet Metab Rep, 2018. 15: p. 15-21.

5. Feinberg, A.P. and B. Tycko, The history of cancer epigenetics. Nature reviews

Cancer, 2004. 4(2): p. 143-53.

6. Zhang, D., et al., Hyperhomocysteinemia promotes inflammatory monocyte

generation and accelerates atherosclerosis in transgenic cystathionine beta-

synthase-deficient mice. Circulation, 2009. 120(19): p. 1893-902.

7. Zhang, D., et al., Severe hyperhomocysteinemia promotes bone marrow-derived

and resident inflammatory monocyte differentiation and atherosclerosis in

LDLr/CBS-deficient mice. Circulation research, 2012. 111(1): p. 37-49.

8. Fang, P., et al., Hyperhomocysteinemia potentiates hyperglycemia-induced

inflammatory monocyte differentiation and atherosclerosis. Diabetes, 2014.

63(12): p. 4275-90.

9. Yang, J., et al., Chronic Kidney Disease Induces Inflammatory CD40+ Monocyte

Differentiation via Homocysteine Elevation and DNA Hypomethylation.

Circulation research, 2016. 119(11): p. 1226-1241.

10. Subramaniam, D., et al., DNA methyltransferases: a novel target for prevention

and therapy. Frontiers in oncology, 2014. 4: p. 80.

11. Kuchiba, A., et al., Global methylation levels in peripheral blood leukocyte DNA

by LUMA and breast cancer: a case-control study in Japanese women. British

journal of cancer, 2014. 110(11): p. 2765-71.

12. Cho, M., et al., Hypomethylation of the MN/CA9 promoter and upregulated

MN/CA9 expression in human renal cell carcinoma. British journal of cancer,

2001. 85(4): p. 563-7.

13. Oshimo, Y., et al., Promoter methylation of cyclin D2 gene in gastric carcinoma.

International journal of oncology, 2003. 23(6): p. 1663-70.

14. Akiyama, Y., et al., Cell-type-specific repression of the maspin gene is disrupted

frequently by demethylation at the promoter region in gastric intestinal

metaplasia and cancer cells. The American journal of pathology, 2003. 163(5): p.

1911-9.

15. Giacinti, C. and A. Giordano, RB and cell cycle progression. Oncogene, 2006.

25(38): p. 5220-7.

16. Hesari, A., et al., Evaluation of the two polymorphisms rs1801133 in MTHFR and

rs10811661 in CDKN2A/B in breast cancer. Journal of cellular biochemistry,

2018.

17. Kaminskas, E., et al., Approval summary: azacitidine for treatment of

myelodysplastic syndrome subtypes. Clinical cancer research : an official journal

of the American Association for Cancer Research, 2005. 11(10): p. 3604-8.

18. Kihslinger, J.E. and L.A. Godley, The use of hypomethylating agents in the

treatment of hematologic malignancies. Leukemia & lymphoma, 2007. 48(9): p.

1676-95.

19. Aavik, E., M. Babu, and S. Yla-Herttuala, DNA methylation processes in

atheosclerotic plaque. Atherosclerosis, 2019. 281: p. 168-179.

20. Hiltunen, M.O., et al., DNA hypomethylation and methyltransferase expression in

atherosclerotic lesions. Vascular medicine (London, England), 2002. 7(1): p. 5-11.

21. Aavik, E., et al., Global DNA methylation analysis of human atherosclerotic

plaques reveals extensive genomic hypomethylation and reactivation at imprinted

locus 14q32 involving induction of a miRNA cluster. European heart journal, 2015.

36(16): p. 993-1000.

22. Deng, Q., et al., Genomic 5-mC contents in peripheral blood leukocytes were

independent protective factors for coronary artery disease with a specific profile

in different leukocyte subtypes. Clinical epigenetics, 2018. 10: p. 9.

23. Jamaluddin, M.D., et al., Homocysteine inhibits endothelial cell growth via DNA

hypomethylation of the cyclin Agene. Blood, 2007. 110(10): p. 3648-55.

24. Zhang, S., et al., m6A Demethylase ALKBH5 Maintains Tumorigenicity of

Glioblastoma Stem-like Cells by Sustaining FOXM1 Expression and Cell

Proliferation Program. Cancer cell, 2017. 31(4): p. 591-606.e6.

25. Liu, J., et al., m6A mRNA methylation regulates AKT activity to promote the

proliferation and tumorigenicity of endometrial cancer. Nature cell biology, 2018.

20(9): p. 1074-1083.

26. Wierda, R.J., et al., Epigenetics in atherosclerosis and inflammation. Journal of

cellular and molecular medicine, 2010. 14(6A): p. 1225-40.

27. Yi, F. and P.-L. Li, Mechanisms of homocysteine-induced glomerular injury and

sclerosis. American journal of nephrology, 2008. 28(2): p. 254-64.

28. Smolle, M. and J.L. Workman, Transcription-associated histone modifications

and cryptic transcription. Biochimica et biophysica acta, 2013. 1829(1): p. 84-97.

29. Shao, Y., et al., Metabolic Diseases Downregulate the Majority of Histone

Modification Enzymes, Making a Few Upregulated Enzymes Novel Therapeutic

Targets--"Sand Out and Gold Stays". Journal of cardiovascular translational

research, 2016. 9(1): p. 49-66.

30. Hancock, J.F., K. Cadwallader, and C.J. Marshall, Methylation and proteolysis

are essential for efficient membrane binding of prenylated p21K-ras(B). The

EMBO journal, 1991. 10(3): p. 641-6.

31. Hancock, J.F., H. Paterson, and C.J. Marshall, A polybasic domain or

palmitoylation is required in addition to the CAAX motif to localize p21ras to the

plasma membrane. Cell, 1990. 63(1): p. 133-9.

32. Wang, H., et al., Inhibition of growth and p21ras methylation in vascular

endothelial cells by homocysteine but not cysteine. The Journal of biological

chemistry, 1997. 272(40): p. 25380-5.

33. Xie, H. and S. Clarke, An enzymatic activity in bovine brain that catalyzes the

reversal of the C-terminal methyl esterification of protein phosphatase 2A.

Biochemical and biophysical research communications, 1994. 203(3): p. 1710-5.

34. Xing, Y., et al., Structural mechanism of demethylation and inactivation of

protein phosphatase 2A. Cell, 2008. 133(1): p. 154-63.

35. DeGrande, S.T., et al., Molecular mechanisms underlying cardiac protein

phosphatase 2A regulation in heart. The Journal of biological chemistry, 2013.

288(2): p. 1032-46.

36. Raposo, A.E. and S.C. Piller, Protein arginine methylation: an emerging

regulator of the cell cycle. Cell division, 2018. 13: p. 3.

37. Boisvert, F.-M., et al., Methylation of MRE11 regulates its nuclear

compartmentalization. Cell cycle (Georgetown, Tex ), 2005. 4(7): p. 981-9.

38. Roberts, D.M., et al., Trimethyllysine and protein function. Effect of methylation

and mutagenesis of lysine 115 of calmodulin on NAD kinase activation. The

Journal of biological chemistry, 1986. 261(4): p. 1491-4.

39. Wang, Y., et al., N6-methyladenosine modification destabilizes developmental

regulators in embryonic stem cells. Nature cell biology, 2014. 16(2): p. 191-8.

40. Roundtree, I.A., et al., Dynamic RNA Modifications in Gene Expression

Regulation. Cell, 2017. 169(7): p. 1187-1200.

41. Cui, Q., et al., m6A RNA Methylation Regulates the Self-Renewal and

Tumorigenesis of Glioblastoma Stem Cells. Cell reports, 2017. 18(11): p. 2622-

2634.

42. Luo, Y., et al., NSun2 Deficiency Protects Endothelium From Inflammation via

mRNA Methylation of ICAM-1. Circulation research, 2016. 118(6): p. 944-56.

43. Yao, Z.M. and D.E. Vance, The active synthesis of phosphatidylcholine is

required for very low density lipoprotein secretion from rat hepatocytes. The

Journal of biological chemistry, 1988. 263(6): p. 2998-3004.

44. Dawaliby, R., et al., Phosphatidylethanolamine Is a Key Regulator of Membrane

Fluidity in Eukaryotic Cells. The Journal of biological chemistry, 2016. 291(7): p.

3658-67.

45. Nishimaki-Mogami, T., K. Suzuki, and A. Takahashi, The role of

phosphatidylethanolamine methylation in the secretion of very low density

lipoproteins by cultured rat hepatocytes: rapid inhibition of

phosphatidylethanolamine methylation by bezafibrate increases the density of

apolipoprotein B48-containing lipoproteins. Biochimica et biophysica acta, 1996.

1304(1): p. 21-31.

46. Dai, J., et al., Metabolism-associated danger signal-induced immune response

and reverse immune checkpoint-activated CD40(+) monocyte differentiation. J

Hematol Oncol, 2017. 10(1): p. 141.

47. Venereau, E., C. Ceriotti, and M.E. Bianchi, DAMPs from Cell Death to New Life.

Front Immunol, 2015. 6: p. 422.

48. Shen, H., et al., Co-signaling receptors regulate T-cell plasticity and immune

tolerance. Frontiers in bioscience (Landmark edition), 2019. 24: p. 96-132.

49. Piechutta, M. and A.S. Berghoff, New emerging targets in cancer immunotherapy:

the role of Cluster of Differentiation 40 (CD40/TNFR5). ESMO open, 2019.

4(Suppl 3): p. e000510.

50. Huber, L.A., K. Pfaller, and I. Vietor, Organelle proteomics: implications for

subcellular fractionation in proteomics. Circulation research, 2003. 92(9): p. 962-

51. Wang, L., et al., Novel extracellular and nuclear caspase-1 and inflammasomes

propagate inflammation and regulate gene expression: a comprehensive database

mining study. Journal of hematology & oncology, 2016. 9(1): p. 122.

52. Xu, K., et al., GATA3, HDAC6, and BCL6 Regulate FOXP3+ Treg Plasticity and

Determine Treg Conversion into Either Novel Antigen-Presenting Cell-Like Treg

or Th1-Treg. Front Immunol, 2018. 9: p. 45.

53. Cueto, R., et al., Identification of homocysteine-suppressive mitochondrial ETC

complex genes and tissue expression profile - Novel hypothesis establishment.

Redox biology, 2018. 17: p. 70-88.

54. Bhatia, P. and N. Singh, Homocysteine excess: delineating the possible

mechanism of neurotoxicity and depression. Fundam Clin Pharmacol, 2015. 29(6):

p. 522-8.

55. Huang, X., et al., Identification of novel pretranslational regulatory mechanisms

for NF-kappaB activation. J Biol Chem, 2013. 288(22): p. 15628-40.

56. Yin, Y., et al., Early hyperlipidemia promotes endothelial activation via a

caspase-1-sirtuin 1 pathway. Arterioscler Thromb Vasc Biol, 2015. 35(4): p. 804-

16.

57. Li, X., et al., IL-35 (Interleukin-35) Suppresses Endothelial Cell Activation by

Inhibiting Mitochondrial Reactive Oxygen Species-Mediated Site-Specific

Acetylation of H3K14 (Histone 3 Lysine 14). Arterioscler Thromb Vasc Biol,

2018. 38(3): p. 599-609.

58. Barski, A., et al., High-resolution profiling of histone methylations in the human

genome. Cell, 2007. 129(4): p. 823-37.

59. Greer, E.L. and Y. Shi, Histone methylation: a dynamic mark in health, disease

and inheritance. Nature reviews Genetics, 2012. 13(5): p. 343-57.

60. Cai, L., et al., Aberrant histone methylation and the effect of Suv39H1 siRNA on

gastric carcinoma. Oncology reports, 2014. 31(6): p. 2593-600.

61. Albert, M. and K. Helin, Histone methyltransferases in cancer. Seminars in cell &

developmental biology, 2010. 21(2): p. 209-20.

62. Wei, S., et al., Histone methylation in DNA repair and clinical practice: new

findings during the past 5-years. Journal of Cancer, 2018. 9(12): p. 2072-2081.

63. Wong, M., P. Polly, and T. Liu, The histone methyltransferase DOT1L:

regulatory functions and a cancer therapy target. American journal of cancer

research, 2015. 5(9): p. 2823-37.

64. Michalak, E.M. and J.E. Visvader, Dysregulation of histone methyltransferases in

breast cancer - Opportunities for new targeted therapies? Molecular oncology,

2016. 10(10): p. 1497-1515.

65. Jorgensen, S., G. Schotta, and C.S. Sorensen, Histone H4 lysine 20 methylation:

key player in epigenetic regulation of genomic integrity. Nucleic acids research,

2013. 41(5): p. 2797-806.

66. Migliori, V., et al., Symmetric dimethylation of H3R2 is a newly identified histone

mark that supports euchromatin maintenance. Nature structural & molecular

biology, 2012. 19(2): p. 136-44.

67. Kirmizis, A., et al., Arginine methylation at histone H3R2 controls deposition of

H3K4 trimethylation. Nature, 2007. 449(7164): p. 928-32.

68. Di Lorenzo, A. and M.T. Bedford, Histone arginine methylation. FEBS letters,

2011. 585(13): p. 2024-31.

69. Wu, J. and W. Xu, Histone H3R17me2a mark recruits human RNA polymerase-

associated factor 1 complex to activate transcription. Proceedings of the National

Academy of Sciences of the United States of America, 2012. 109(15): p. 5675-80.

70. Wang, J., et al., Asymmetric Expression of LincGET Biases Cell Fate in Two-Cell

Mouse Embryos. Cell, 2018. 175(7): p. 1887-1901.e18.

71. Perez-Miguelsanz, J., et al., Betaine homocysteine S-methyltransferase emerges

as a new player of the nuclear methionine cycle. Biochimica et biophysica acta

Molecular cell research, 2017. 1864(7): p. 1165-1182.

72. Chen, H., et al., Identification of a homocysteine receptor in the peripheral

endothelium and its role in proliferation. J Vasc Surg, 2005. 41(5): p. 853-60.

73. Jamaluddin, M.S., X. Yang, and H. Wang, Hyperhomocysteinemia, DNA

methylation and vascular disease. Clinical chemistry and laboratory medicine,

2007. 45(12): p. 1660-6.

74. Wang, H., et al., Hyperhomocysteinemia accelerates atherosclerosis in

cystathionine beta-synthase and apolipoprotein E double knock-out mice with and

without dietary perturbation. Blood, 2003. 101(10): p. 3901-7.

75. Cheng, Z., et al., Hyperhomocysteinemia and hyperglycemia induce and

potentiate endothelial dysfunction via mu-calpain activation. Diabetes, 2015.

64(3): p. 947-59.

76. Wang, J.X. and R.R. Breaker, Riboswitches that sense S-adenosylmethionine and

S-adenosylhomocysteine. Biochemistry and cell biology = Biochimie et biologie

cellulaire, 2008. 86(2): p. 157-68.

77. Yang, X.F., et al., Processing sites are different in the generation of HLA-A2.1-

restricted, T cell reactive tumor antigen epitopes and viral epitopes. International

journal of immunopathology and pharmacology, 2006. 19(4): p. 853-70.

78. Yin, Y., et al., Inflammasomes are differentially expressed in cardiovascular and

other tissues. International journal of immunopathology and pharmacology, 2009.

22(2): p. 311-22.

79. Li, X., et al., IL-35 is a novel responsive anti-inflammatory cytokine--a new

system of categorizing anti-inflammatory cytokines. PloS one, 2012. 7(3): p.

e33628.

80. Mudd, S.H., Hypermethioninemias of genetic and non-genetic origin: A review.

American journal of medical genetics Part C, Seminars in medical genetics, 2011.

157C(1): p. 3-32.