The Dysregulation of Polyamine Metabolism in Colorectal Cancer Is Associated with Overexpression of c-Myc and C/EBPβ rather than Enterotoxigenic Bacteroides fragilis Infection

Colorectal cancer is one of the most common cancers in the world. It is well known that the chronic inflammation can promote the progression of colorectal cancer (CRC). Recently, a number of studies revealed a potential association between colorectal inflammation, cancer progression, and infection caused by enterotoxigenic Bacteroides fragilis (ETBF). Bacterial enterotoxin activates spermine oxidase (SMO), which produces spermidine and H2O2 as byproducts of polyamine catabolism, which, in turn, enhances inflammation and tissue injury. Using qPCR analysis, we estimated the expression of SMOX gene and ETBF colonization in CRC patients. We found no statistically significant associations between them. Then we selected genes involved in polyamine metabolism, metabolic reprogramming, and inflammation regulation and estimated their expression in CRC. We observed overexpression of SMOX, ODC1, SRM, SMS, MTAP, c-Myc, C/EBPβ (CREBP), and other genes. We found that two mediators of metabolic reprogramming, inflammation, and cell proliferation c-Myc and C/EBPβ may serve as regulators of polyamine metabolism genes (SMOX, AZIN1, MTAP, SRM, ODC1, AMD1, and AGMAT) as they are overexpressed in tumors, have binding site according to ENCODE ChIP-Seq data, and demonstrate strong coexpression with their targets. Thus, increased polyamine metabolism in CRC could be driven by c-Myc and C/EBPβ rather than ETBF infection.


Introduction
Colorectal cancer (CRC) is one of the most common cancers in the world. It has been estimated that in 2012 about 1.4 million people were diagnosed and more than 690 thousand died [1]. The lifetime risk of developing colorectal cancer is about 5% worldwide. Women have a higher risk for colon cancer that men [2]. The prognosis of colorectal cancer is closely related to the stage of disease at diagnosis [3]. CRC can have no symptoms in early stages, and mean 5-year survival rate for peoples detected at an early stage is about 90% compared to 10% for people diagnosed for cancer with distant metastases [4]. Certain factors increase a risk of developing the disease. These are age [5], polyps of the colon [6], history of cancer, heredity [7][8][9], smoking [10], diet and microbiota [11,12], lack of physical activity [13], chronic inflammation (colitis and IBD) [14,15], viruses [16,17], and exogenous hormones [18]. The chronic inflammation caused by infection is one more risk factor for colorectal cancer [19,20]. Some pathogenic strains of Escherichia coli (cyclomodulin-positive) 2 Oxidative Medicine and Cellular Longevity are able to induce chronic inflammation and can be involved in carcinogenesis. Cyclomodulin-positive E. coli strains were more prevalent in both the mucosa and tumors of patients with colorectal cancer (26% patients) versus diverticulosis control (6% patients) [19]. In addition, the number of colonic polyps was elevated in multiple intestinal neoplasia (Min) mice inoculated with a colon cancer-associated E. coli strain (11G5) [19].
Recent studies have demonstrated that the enterotoxigenic Bacteroides fragilis (ETBF) bacterium is an important cause of chronic inflammation in human and animal colon. It has been presented that the bft gene, which encodes Bacteroides fragilis toxin (BFT), is associated with colorectal neoplasia and may be a risk factor for developing CRC [21]. It was observed that BFT produced by bacteria upregulates both spermine oxidase (SMOX) gene expression at mRNA and protein levels in cultures of human normal colonic epithelial cells [20]. SMOX encodes SMO protein, which plays an important role in the regulation of polyamine metabolism. SMO catalyzes the oxidation of spermine to spermidine and produces hydrogen peroxide (H 2 O 2 ) and aldehydes [22]. This results in apoptosis, DNA damage, and consequently the development of cancer. For example, cytotoxin produced by Helicobacter pylori strains causes an increase in spermine oxidase levels in human gastric epithelial cells. These pathogenic H. pylori strains contain cytotoxin-associated gene A (CagA) and represent a risk factor for gastric cancer. The strong association of H. pylori (Cag+) infection, SMO levels, apoptosis, and oxidative DNA damage has been observed [23,24].
In recent years, we have seen a steady increase in the number of studies examining the role of intracellular polyamine metabolism in tumor development. Several important enzymes, spermidine/spermine N1-acetyltransferase (SSAT), N1-acetylpolyamine oxidase (APAO), and SMO, appear to play critical roles in many cancers. All such enzymes are highly inducible by multiple stress signals, including ones caused with bacterial pathogens, and have the potential to alter polyamine homeostasis. APAO and SMO enzymes produce reactive oxygen species (ROS), H 2 O 2 , and aldehydes, which are potentially harmful to cells. ROS are key signaling molecules, which play an important role in several pathways (e.g., NF-B, ERK1/2, p38, PI3K, and others) and can contribute to the induction of inflammation and cancer [25].
Thus, the abundant pathogenic microbiota alters the host tissue microenvironment leading to chronic inflammation, immune dysregulation, and elevated levels of ROS. All these may result in activation of oncogenes, downregulation of tumor suppressor genes, DNA damage, and cell and tissue injury, thereby contributing to tumor growth. In the colon, the alteration of polyamine catabolism caused by infection with consequent H 2 O 2 generation and DNA damage may be a common cause of inflammation and promotion of carcinogenesis. Moreover, an increase in polyamine catabolism rates and the production of H 2 O 2 has been involved in the response to chemotherapeutic agents or specific antitumor polyamine analogues in several tumors, including colorectal cancer [26][27][28][29][30]. However, the clinical, molecular, and prognostic associations of infection and the expression of polyamine metabolism gene in colorectal cancer remain unclear.

Tissue Specimens.
A total of 50 paired specimens of stages I-IV colorectal cancer (CRC) and adjacent morphologically normal tissues were taken from patients with primary carcinoma of the colon and rectum, which had not been exposed to radiation or chemotherapy, during surgical resection. Each sample was frozen and placed in liquid nitrogen immediately after surgery. The specimens were characterized according to the American Joint Committee on Cancer (AJCC) staging system [31]. The diagnosis was verified by histopathology and only samples containing 70-80% or more tumor cells were used in the study. The tissue samples were collected in accordance with the guidelines issued by the Ethics Committee, National Medical Research Radiological Center, the Ministry of Health of the Russian Federation. All patients gave written informed consent, which is available upon request. The study was carried out in accordance with the principles outlined in the Declaration of Helsinki (1964).

RNA and DNA Isolation and cDNA Synthesis.
Total RNA was isolated using Micro-Dismembrator S (Sartorius, Germany) and RNeasy Mini Kit (Qiagen, Germany) in accordance with the manufacturer's instructions. For the detection of bacteria, DNA was extracted using the QIAamp DNA Mini Kit (Qiagen, Germany) and further treated with proteinase K in accordance with the manufacturer's protocol. Purified RNA and DNA were quantified using Qubit 2.0 fluorometer (Invitrogen, USA) and their quality was determined by Agilent Bioanalyzer 2100 (Agilent Technologies, USA). All RNA samples were treated with DNase I (Thermo Fisher Scientific, USA), and cDNA was synthesized using M-MLV Reverse Transcriptase (Thermo Fisher Scientific, USA) and random hexamers according to standard manufacturer's protocol.

qPCR.
To detect and quantitate Bacteroides fragilis we used the primers targeting Bft1 gene. These primers were taken from the work of Viljoen et al. [32]: forward 5 -GACGGTGTATGTGATTTGTCTGAGAGA-3 , reverse 5 -ATCCCTAAGATTTTATTATCCCAAGTA-3 . EvaGreen Dye (Biotium Inc., USA) was used as fluorescent DNAbinding dye for the detection and quantification of PCR products. Purified bacterial control DNA was obtained from Orekhovich Institute of Biomedical Chemistry, Russian Academy of Sciences (Moscow).
To evaluate gene expression, we used TaqMan Gene Expression Assays (Thermo Fisher Scientific, USA). This consists of a pair of specific unlabeled PCR primers and a TaqMan probe with a FAM dye label on the 5 end and minor groove binder (MGB) nonfluorescent quencher (NFQ) on the 3 end. All probes contained the dye FAM at 5 -end and RTQ1 at 3 -end. qRT-PCR was performed as described earlier using RPN1 and GUSB reference genes [33][34][35].
Oxidative Medicine and Cellular Longevity 3 All reactions were performed using AB 7500 Real-Time PCR System (Thermo Fisher Scientific, USA) with RQ (Relative Quantitation) software (Thermo Fisher Scientific, USA). PCR program was as follows: 10 min at 95 ∘ C and then 50 twostep cycles 15 s at 95 ∘ C and 60 s at 60 ∘ C. The total reaction volume was 20 L in triplicate. PCR products were analyzed in 2% agarose gels, and nucleotide sequences of the amplicons were verified by Sanger sequencing with ABI Prism 3100 Genetic Analyzer (Thermo Fisher Scientific, USA).

Analysis of qRT-PCR Data.
For the detection and quantification of bacterial DNA, absolute quantification method was used. A standard curve was constructed using serial 10fold dilutions of control bacterial DNA. The genome size of Bacteroides fragilis (5.3 Mb) and the mass of DNA per genome were used to calculate the concentration of bacterial DNA [36]. Spearman's rank correlation analysis was used to check the dependence between target gene expression levels and the concentration of bacterial DNA within the same samples.
mRNA qRT-PCR data were analyzed using the relative quantification method (ΔΔC t ) taking into account the efficiency of the PCR amplification using ATG (Analysis of Transcription of Genes) tool as described in [34,37]. The relative inner variability between mRNA levels of reference genes (RPN1 and GUSB) was not higher than twofold in tumor and normal tissues; therefore, twofold and higher mRNA level changes for the target genes were considered as significant. We used nonparametric Mann-Whitney test to validate the significance of gene expression alterations (p ≤ 0.01 was taken as the criterion of statistical significance). All statistical analyses were performed in the R environment. Pearson correlation coefficient was used to evaluate the coexpression of human genes (p ≤ 0.001 was taken as the criterion of statistical significance). We supplemented the coexpression analysis with ENCODE ChIP-Seq data using previously developed CrossHub tool [38].

Expression of SMOX Gene and Bacteroides fragilis Quantification.
We analyzed 36 paired samples of primary colorectal carcinomas and adjacent normal tissues to quantify ETBF. Serial dilution of genomic DNA from ETBF was used as standard. At least one copy of ETBF DNA per 50 ng of total extracted DNA was detected in 8 paired tumor and adjacent normal samples, including three samples, which showed a significantly increased amount of the ETBF DNA (>1000 copies/1 ng of extracted DNA). Most of the tested samples (23/36) demonstrated less than a copy of bacterial DNA per 50 ng of total extracted DNA. Five samples were found to be ETBF-negative ( Figure 1).

Expression Quantification of Genes Involved in Polyamine
Metabolism and Regulation of Inflammatory Response. The set of 36 colorectal tumors and adjacent normal tissues was extended with an additional set of 14 paired samples in order to increase the statistical significance of the results. Thus, we used a set of 50 colorectal cancers to evaluate the expression of target genes, which encode enzymes of polyamine metabolism and genes participating in immune response and inflammation signaling.
Using RT-qPCR, we evaluated the relative expression level of 17 genes involved in polyamine metabolism and 4 genes mediating metabolic reprogramming, and the regulation of inflammatory response and cell proliferation: c-Myc (MYC), n-Myc (MYCN), Max, and C/EBP gene encoding CCAAT/enhancer binding protein (CEBPB). The results are given in Figure 2

Discussion
Bacteroides fragilis contains up to 1%-2% of the normal colonic microbial flora in humans [39,40]. Pathogenic strains of B. fragilis that produce enterotoxin are associated with the development of inflammatory diarrheal disease in both children and adults [41,42], IBD [43], and colitis [44], thereby contributing to chronic inflammation and onset of colon tumors. It was supposed that B. fragilis toxin (BFT), produced only by enterotoxigenic strains (ETBF), binds to a specific intestinal epithelial cell receptor and induces several signal transduction pathways [45,46]. BFT cleaves E-cadherin [47,48] and activates Wnt/ -catenin pathway signaling [49]. BFT was also shown to be involved in activation of tyrosine kinases, MAPKs, and NF-B signaling pathways and to be able to increase cellular proliferation mediated by the elevated expression of c-Myc oncogene [50][51][52][53]. Moreover, as mentioned above, BFT may be implicated in polyamine catabolism through upregulation of SMO protein levels and enzyme activity [20]. Recent studies suggested SMO enzyme as a potential source of an inflammation-associated ROS produced during polyamine catabolism [54]. Thus, the association between ETBF colonization rates of CRC patients and the expression of SMOX gene appears to be an important link between chronic inflammation caused by infection, tumor onset, and progression.
In this study, we quantitated ETBF strain colonization rates and evaluated the expression of SMOX gene at mRNA level in CRC patients using qPCR. We found that the majority of CRC patients were colonized with toxin-producing strains of B. fragilis, but only small amounts of bacterial DNA were identified. A significantly increased content of bacterial DNA  Stage II  Stage III  Stage IV   T2N0M0  T2N0M0  T2N0M0  T2N0Mx  T3N0Mx  T3N0M0  T3N0M0  T3N0Mx  T3N0M0  T3N0M0  T3N0M0  T3N0M0  T3N0M0  T3N0Mx  T2N1M0  T3N2M0 T3N1Mx  T3N1M0  T4N1Mx  T3N2Mx  T2N2M0  T3N1Mx  T3N2M0  T3N1M0  T4N1M1  T4N2M1  T3N0M1  T3N2M1  T4N1M1  T3N2M1  T4N0M1  T3N1M1  T4N0M1  T3N1M1  T4N0M1  was detected in three patients with I and II stages of CRC (in both tumor and normal tissues). In contrast, pathogenic cyclomodulin-positive E. coli strains, a possible cofactor of colorectal carcinogenesis, were predominantly found on mucosa of patients with stages III/IV [19]. Additional studies and extended sampling are needed to reveal possible associations between ETBF colonization and disease stage. The expression of SMOX gene revealed no statistically significant correlation with the amount of bacterial DNA, but the samples with a high concentration of bacterial DNA (>1000 copies per 50 ng) demonstrated high SMOX expression levels in both normal tissue and tumor (relatively to GUSB and RPN1 reference genes, Figure 1). These results do not contradict the hypothesis that ETBF strains can be proinflammatory and oncogenic bacteria, but additional studies are needed to understand the possible role of ETBF in colorectal carcinogenesis [21]. These results are consistent with a mechanism of SMOX induction independent of ETBF infection.
Next, we tested differential expression of 17 genes participating in polyamine metabolism and 4 genes involved in the mediation of metabolic preprogramming, cell proliferation (c-Myc, n-Myc, and Max) and inflammation (C/EBP ). c-Myc and n-Myc form dimers with Max, translocate to the nucleus, and then activate the transcription of many genes participating in cell cycle regulation, glycolysis, energy metabolism, hypoxic adaptation, DNA replication, and other processes [55][56][57][58]. C/EBP is a transcription factor, which can form either homodimers or heterodimers with other CCAAT/enhancer binding proteins (alpha, delta, and gamma). C/EBP is known to be a mediator of inflammation and immunity [59,60]. The most important C/EBP targets are interleukins IL-6, IL-4, IL-5, and TNF- [61][62][63][64]. We found upregulation of c-Myc, n-Myc, and C/EBP in the analyzed samples. However, we did not find overexpression of Max. This is in agreement with a previous finding: overexpression of c-Myc and n-Myc but not Max is observed in many tumors [55,58,65,66]. Our data showed increased expression of SMOX gene at all stages of colorectal cancer, this tendency being more pronounced at the early stages ( Figure 2). Spearman correlation coefficient between disease stage and tumor/normal fold change of SMOX mRNA level is = −0.19: SMOX expression tends to be lower with disease stage. However, this tendency is not statistically Oxidative Medicine and Cellular Longevity 5   T2N0M0  T2N0M0  T2N0Mx  T2N0M0  T2N0M0  T2N0M0   T3N0Mx  T3N0M0  T3N0M0  T3N0M0  T3N0M0  T4N0M0  T3N0M0  T3N0M0  T3N0M0  T3N0M0  T3N0M0  T3N0Mx  T2N1M0  T3N0M0  T3N0M0  T3N0M0  T3N0M0  T3N0M0   T3N2M0  T3N1Mx  T3N1M0  T4N1Mx  T3N2Mx  T2N2M0  T3N2M0  T3N1M0   T4N1M1  T4N2M1  T3N0M1  T3N2M1  T4N1M1  T4N1M1  T3N2M1  T3N2M1  T4N0M1  T3N1M1  T4N0M1  T3N1M1  T4N1M1  T4N0M1  T3N1M1  T4N1M1  T3N2M1  Expression level fold change (T/N) * * * significant ( > 0.05). Different tumors demonstrate dramatically distinct SMOX expression profiles. SMO was found to be elevated in both prostate adenocarcinoma and prostatic intraepithelial neoplasia [67] whereas breast cancer showed SMOX underexpression at both mRNA and protein levels [68]. Other genes that showed statistically significant correlation with disease stage were MTAP ( = −0.30, = 0.03), and SRM ( = −0.27, = 0.05), which tended to decrease the expression level with tumor stage, and OAZ3 ( = 0.42, = 0.003), which increased its expression with disease progression.
Polyamines are organic cations, which are essential for cell proliferation and growth, and their levels are frequently altered in many human tumors [69,70]  also been shown to play an important role in inflammationinduced carcinogenesis [71]. Intracellular polyamine metabolism occurs via two pathways: classical and alternate ( Figure 4). Polyamines (putrescine, spermidine, and spermine) are synthesized by mammalian cells, while agmatine is produced only by plants and bacteria, including intestinal microbial flora [72][73][74]. L-arginine carboxylase (ADC), found only in nonmammals, catalyzes a reaction of decarboxylation of L-arginine to agmatine [72,75]. The latter is further hydrolyzed to putrescine and urea by agmatinase, which is encoded by AGMAT gene. Recent reports also supposed that different pathogens, such as viruses and bacteria, could upregulate agmatinase at mRNA and protein levels, thereby inducing polyamine synthesis [76,77].
In the present study, we revealed that the expression of ADC gene decreased on the average by half in CRC samples compared to adjacent normal tissues. These findings indicate that, in case of colorectal cancer, the enhanced polyamine synthesis is not associated with alternate path and the intestinal microbiota does not significantly contribute to these processes.
The three major enzymes that are involved in polyamine metabolism (SSAT, APAO, and SMO) are encoded by SAT1, PAOX, and SMOX genes, respectively. One of the products of both SMO and APAO-mediated reactions are H 2 O 2 and highly toxic aldehydes, which may also cause oxidative damage. SSAT cellular toxicity is thought to be caused by produced acetylated polyamines, which are utilized by APAO. APAO is constitutively expressed, while SSAT is an inducible enzyme [78]. SSAT expression is induced by different stimuli, such as toxins, hormones, cytokines, nonsteroidal antiinflammatory agents, natural products, and pathogens. It is regulated via several pathways including TNF-and NF-B [79][80][81][82]. SSAT expression can be also mediated by Nrf2 in response to the addition of H 2 O 2 in human hepatoma HUH7 cells [83]. We found that SAT1 gene was differentially expressed in many CRC samples (either up-or downregulated), while a decreased expression of PAOX gene was observed. This indicates that the SSAT/APAO pathway is not hyperactive and does not seem to be a cause of oxidative damage in colorectal cancer. Inflammatory response agents and stress pathways are not the result of the induction of SSAT expression at mRNA level. Thus, it is possible that the increased SMOX gene expression and enzyme activity make the greatest contribution to the oxidative stress damage caused by polyamine catabolism in colorectal cancer [20].
It is known that specific protein products of oncogenes and tumor suppressor genes can regulate polyamine metabolism [84][85][86]. The results of coexpression analysis coupled to ENCODE ChIP-Seq data strongly suggest c-Myc and C/EBP as regulators of the expression of key enzymes of polyamine metabolism that are upregulated in colorectal cancer: SMOX, AZIN1, MTAP, SRM, AMD1, ODC1, and AGMAT. It should be mentioned that genes encoding polyamine metabolic enzymes are regulated transcriptionally/post-transcriptionally by changes in the levels of intracellular polyamines. Additional studies (c-Myc and C/EBP knockdown assays) are needed to prove our finding. Not surprisingly, three of these genes, ODC1, AMD1, and SRM, are already known c-Myc targets [87][88][89]. Besides these genes, c-Myc is also known to regulate eIF5A2. However, this gene did not demonstrate coexpression with c-Myc but did show ChIP-Seq association (c-Myc binding site). It is known that c-Myc induces the expression of ornithine decarboxylase (ODC), which catalyzes the first rate-limiting step in polyamine synthesis [89]. It has been also shown that bacterial infections can stimulate polyamine synthesis through ODC [90], whereas suppression of ODC leads to the depletion of cellular polyamine levels in human colorectal cancer cells [82,91]. ODC inhibition with difluoromethylornithine (DFMO) is a possible anticancer therapy.  (2) putrescine is formed from the reaction of ornithine decarboxylation catalyzed by ODC1 (ornithine decarboxylase-1). OAZ can bind to ODC1 to form OAZ-ODC1 complex and subsequently reduce polyamine synthesis. AZIN1 (antizyme inhibitor-1) brakes the ODC1-OAZ complex and liberates ODC1; (3) AMD1 (S-adenosylmethionine decarboxylase) decarboxylates Sadenosylmethionine (SAM) to decarboxylated SAM (dcSAM); (4) dcSAM provides aminopropyl groups to putrescine to produce spermidine by spermidine synthase (SRM) and spermine by spermine synthase (SMS). MTA (methylthioadenosine) is generated as a byproduct. Spermine can be recycled back to spermidine directly by spermine oxidase (SMOX). Spermine and spermidine can be recycled to spermidine and putrescine by spermidine/spermine-N1-acetyltransferase (SAT1) followed by oxidation by polyamine oxidase (PAOX) [101]. MTA can be processed to the methionine: MTA phosphorylase (MTAP) catalyzes the cleavage of MTA yielding 5-methylthioribose-1-phosphate (MTRu-P), which is further metabolized to DHKMP (1,2-dihydro-3-keto-5-methylthiopentene). ADI (acireductone dioxygenase) catalyzes DHKMP to 2-oxo-4-methylthiobutyrate (KMTB) and transamination of KMTB results in formation of methionine [102][103][104]. polyps in patients with history of resected sporadic colorectal adenomas [92].
n-Myc (MYCN) is oncogenic transcription factor, which can directly upregulate ODC expression in neuroblastomas [93]. Moreover, it has been shown that the reduction of n-Myc protein levels through inhibition of glycolysis may decrease ODC expression and potentiate polyamine levels in human neuroblastoma cell lines [84]. We showed that expression of ODC1 gene was increased in a number of CRC samples, as well as mRNA level of c-Myc and n-Myc genes. However, n-Myc showed no expression correlations with ODC1 and no ENCODE ChIP-Seq data are available for n-Myc.
It is worthy of note that the expression of OAZ1, OAZ2, and OAZ3 genes, which encode major regulators of ornithine decarboxylation to putrescine through inhibition of ODC catalytic activity, and the expression of AZIN1 gene, which is involved in inhibition of antizyme (OAZ) family, did not change in the majority of CRC samples compared with normal samples. This indicates that the production of putrescine, catalyzed by ODC during polyamine catabolism, is not suppressed in colorectal cancer.
We have observed a significant increase in eIF5A2 mRNA levels in colorectal tumors. Cytosolic protein encoded by eIF5A2 gene undergoes posttranslational modification of Lys 50 to hypusine [94]. Spermidine as a substrate is involved in the first step of this process. The eIF5A2 protein is essential for eukaryotic cell proliferation, but the molecular function of eIF5A remains incompletely clear. It was shown that c-Myc can possibly transactivate the eIF5A2 gene [95,96]. eIF5A2, in turn, was demonstrated to regulate MTA1 (metastasis-associated 1) via c-Myc in gastric cancer and colorectal carcinoma [97,98]. Our data suggests that eIF5A2 can play an important oncogenic role in CRC and enhance the involvement of polyamines in this pathological process.
We elevated the expression of C/EBP gene, an important transcription factor, which controls the expression of genes involved in inflammatory response [60,99,100]. We observed a positive correlation between C/EBP gene expression and many key genes such as ADC, SMOX, AGMAT, and SRM involved in polyamine metabolism.

Conclusions
In summary, our results show that dysregulation of polyamine metabolism in all stages of CRC can be associated with chronic inflammation mediators rather than with the infection caused by ETBF. This is the first report that presents changed levels of expression of the key components of polyamine metabolism. We characterized some important aspects of the expression of SMOX and PAOX genes, which are responsible for cellular ROS generation. Two transcription factors, oncogenic c-Myc (responsible for metabolic reprogramming and cell proliferation) and C/EBP (mediator of inflammation and immune response), were found to be the most likely regulators of several key enzymes of polyamine metabolic pathway.