Chapter 3: Phenotypic Characterisation of Melanotransferin Knockout Mouse and Down-Regulation of Melanotransferrin in Melanoma Cells

CHAPTER 3: PHENOTYPIC CHARACTERISATION OF MELANOTRANSFERIN KNOCKOUT MOUSE AND DOWN-REGULATION OF MELANOTRANSFERRIN IN MELANOMA CELLS

“An ignorant person is one who doesn't know what you have just found out.” Will Rogers, 1879 - 1935

This Chapter is published in:

1. Sekyere, E.O., Dunn, L.L., Suryo Rahmanto, Y., and Richardson, D.R. (2006) Role of melanotransferrin in iron metabolism: studies using targeted gene disruption in vivo. Blood. 107(7):2599-601. IF 2006: 10.4 *Selected by the Editors of Blood as a Plenary Paper which this journal describes as “papers of exceptional scientific importance”.

2. Dunn, L.L., Sekyere, E.O., Suryo Rahmanto, Y., and Richardson, D.R. (2006) The function of melanotransferrin: a role in melanoma cell proliferation and tumorigenesis. Carcinogenesis. 27(11):2157-69. IF 2006: 5.3

3.1. Introduction

Melanotransferrin or melanoma tumour antigen p97 is an iron (Fe)-binding transferrin

(Tf) homologue originally identified at high levels in melanomas and other tumours, cell lines and foetal tissues [Brown et al. 1981a; Brown et al. 1982; Woodbury et al.

1980]. Initial studies found MTf to be absent or only slightly expressed in normal adult tissues [Brown et al. 1981a], while later investigations demonstrated MTf in a range of normal tissues [Alemany et al. 1993; Rothenberger et al. 1996; Sciot et al. 1989].

Recent studies showed MTf to be expressed at higher levels in the brain and epithelial surfaces of the salivary gland, pancreas, testis, kidney and sweat gland ducts compared with other normal tissues [Richardson 2000; Sekyere et al. 2005; Sekyere et al. 2006].

Nonetheless, the highest levels of MTf are still found in melanoma cells, in particular the SK-Mel-28 melanoma cell line which expresses 3.0-3.8 x 105 MTf molecules per cell [Brown et al. 1981a; Brown et al. 1982].

The MTf molecule shares many properties in common with the Tf family of proteins, including the ability of isolated and purified MTf to bind Fe from Fe(III) citrate complexes [Baker et al. 1992; Brown et al. 1981b; Brown et al. 1982; Plowman et al.

1983; Rose et al. 1986]. These characteristics and the high expression of MTf in melanoma cells suggested that it played a role in Fe transport, possibly assisting these tumour cells with their increased Fe requirements [Richardson & Baker 1990;

Richardson & Baker 1991b; Richardson 2000]. However, unlike Tf, MTf is bound to the membrane by a glycosyl-phosphatidylinositol anchor [Sekyere & Richardson 2000].

Previous in vitro studies demonstrated that, although MTf could bind Fe, it did not

efficiently donate it to the cell [Richardson & Baker 1991a; Richardson & Baker 1990;

Richardson & Baker 1991b; Richardson 2000].

To date, the function of MTf in both normal and neoplastic cells remains unknown.

Many roles have been proposed for MTf involvement in physiological and pathological processes, including transcytosis of Fe across the blood brain barrier, angiogenesis, cell migration, plasminogen activation, chondrogenesis, eosinophil differentiation and

Alzheimer’s disease [Demeule et al. 2002; Moroo et al. 2003; Sala et al. 2002].

However, definitive evidence for the functional role of MTf is lacking.

Melanotransferrin knockout (MTf -/-) mice were generated by targeted gene disruption to determine the role of MTf. This Chapter provides a comprehensive assessment of the phenotype of this animal which is important for defining the function of MTf. This investigation shows that lack of MTf expression had no significant effect on the growth, development, behaviour, or metabolism of MTf -/- mice compared with wild-type

(MTf +/+) littermates. However, whole genome microarray analysis showed changes in the expression of genes such as myocyte enhancer factor 2a (Mef2a), transcription factor 4 (Tcf4), glutaminase (Gls) and apolipoprotein D (Apod) when comparing MTf -/- mice with MTf +/+ littermates.

To assess the role of MTf in neoplastic cells, MTf expression was down-regulated by post-transcriptional gene silencing (PTGS) in SK-Mel-28 melanoma cells. This study also shows that down-regulation of MTf expression impaired cellular proliferation,

DNA synthesis, cellular migration in vitro and tumour formation in vivo.

Collectively, the work described in this Chapter is the first to conclusively demonstrate that MTf has no role in Fe metabolism in normal and neoplastic cells.

3.2. Materials and Methods

3.2.1. Animals

Animal work was conducted in accordance with the University of New South Wales

Animal Ethics Committee Guidelines. All mice were housed under a 12 h light-dark cycle, fed routinely with basal rodent chow (0.02% Fe) and watered ad libitum.

3.2.2. Generation of MTf -/- mice

MTf -/- mice were generated by homologous gene targeting in embryonic stem (ES) cells. The targeting vector was linearised and electroporated into 129/SvJ ES cells grown in medium containing G418 (Invitrogen). The ES cell clones targeted at both the

5’ and 3’ ends were used to generate chimeras.

High-percentage chimeric males were mated with C57BL/6 females to obtain germ line–transmitting heterozygous offspring of mixed 129/SvJ x C57BL/6 background. To delete the neor-cassette, male heterozygote (MTf +/- neo+) mice were mated to B6-deleter females to obtain (MTf +/- cre+ neo-) offspring. MTf +/- cre+ females were crossed with

C57BL/6 males to remove cre, generating MTf +/- mice free of neor and cre. These heterozygotes were intercrossed to generate MTf +/+ and MTf -/- littermates.

3.2.3. Isolation and Genotyping of Genomic DNA from Animal Tissues

Three week old mice were tail-tipped (approximately 5 mm of tail per mouse) and the tissue placed in sterile tubes. The tissue was dissolved overnight at 55°C and genomic

DNA isolated and purified for PCR analysis using the DNeasy® Tissue Kit (Qiagen) in accordance with the protocol provided by the manufacturer.

Genotyping was carried out by Southern analysis [Sambrook et al. 1989] and polymerase chain reaction (PCR) using genomic tail DNA. Briefly, the PCR reaction mixture contained 200 μM of each dNTP, 0.5 μM of each primer, 1.5 mM MgCl2, 1x

PCR reaction mixture and 1.0 U Taq DNA polymerase (Invitrogen). The PCR cycles were consisted of 3 min at 94°C, followed by 30-40 cycles at 94°C (30 s), 55-60°C (1 min) and 72°C (1 min), with a single final extension period of 5 min at 72°C. The PCR product was analysed on a 1.5% agarose gel in 1x TAE buffer.

3.2.4. Serum chemistry, haematology and histochemistry

Serum chemistry and haematological parameters were determined using a Konelab 20i analyser (Thermo-Electron Corporation, Finland) and Sysmex K-4500 analyser (TOA

Medical Electronics Co., Kobe, Japan), respectively. The total Fe-binding capacity

(TIBC) was calculated by adding the serum Fe level and the unbound Fe-binding capacity (UIBC). The Tf saturation was calculated as serum Fe level/TIBC x 100.

Blood and bone marrow smears were prepared and stained with Giemsa. A wide variety of tissues were dissected, fixed in formalin, sectioned and stained with haematoxylin and eosin and Prussian blue. These were assessed by two independent veterinary pathologists.

3.2.5. Total tissue iron, copper and zinc determinations

Total tissue Fe, copper (Cu) and zinc (Zn) concentrations were measured using inductively coupled plasma atomic emission spectrometry (ICP-AES) via standard techniques [Kingston & Jassie 1988].

3.2.6. Microarray processing and analysis

Six 17.5 week old female mice comprising four MTf -/- and two MTf +/+ from the same litter were used. Total RNA was isolated by grinding brain tissue in 1 mL of TRIzol® reagent (Invitrogen, Sydney, Australia). First strand cDNA synthesis was performed using a total of 15 μg of RNA by the Affymetrix One Cycle® cDNA synthesis kit and the cDNA products purified using the Affymetrix GeneChip® sample clean-up kit

(Millenium Sciences, Victoria, Australia). Using the Affymetrix IVT® labelling kit

(Millenium Sciences), biotin-labeled cRNA was prepared from the cDNA template.

After further purification with the above mentioned clean-up kit, the quantity of product was ascertained using the Nanochip protocol on an Agilent Bioanalyser 2100 (Agilent

Technologies, CA, USA).

A total of 20 μg of labeled cRNA was fragmented to the 50-200 bp size range and the quality of the fragments checked again. Samples (cRNA, 0.05 μg/μL) that passed this checkpoint were then prepared for hybridisation to the mouse GeneChip® 430 2.0. The mouse GeneChip® 430 2.0 consists of greater than 39,000 transcripts and variants from over 34,000 well characterised mouse genes. On completion of hybridisation and washing, microarray chips were scanned with the Affymetrix GeneChip® Scanner 3000

(Millenium Sciences). Microarray data is available on the Gene Expression Omnibus

(http://www.ncbi.nlm.nih.gov/geo/) using the accession numbers GSM101412,

GSM101413, GSM101414, GSM101415, GSM101416 and GSM101417.

A two phase strategy was used to identify differentially expressed genes. First, genome- wide screening was performed using Affymetrix GeneChips®. The empirical Bayes procedure [Lonnstedt & Speed 2002] was applied in order to detect genes most likely to be differentially expressed between the MTf -/- and MTf +/+ samples. Individual p-values were then adjusted using the Holm step-down procedure to reduce the likelihood of false positives [Holm 1979]. Statistical analysis of data from the Affymetrix

Genechips® was used to produce a list of thirty genes with the greatest fold change. This analysis was not meant to provide proof of differential expression. Rather, definitive evidence of differential expression was obtained from the RT-PCR assessment of samples used for the microarray analysis and also at least three other independent samples.

3.2.7. Construction of siRNA vectors, cell culture and transfections

To prepare a construct capable of generating hairpin siRNA specific for human MTf mRNA, the pSilencer™ 3.1-H1 neo expression vector (Ambion, Texas, USA) was used.

The vector was used to clone four separate transgenes (A-D) of 66-bp each that transcribe 19-mer double-stranded hairpin RNAs of the target gene. The four transgenes were specifically targeted to positions 2046-2064-bp (A), 2031-2049-bp (B), 2048-

2066-bp (C) and 1416-1434-bp (D) in the MTf gene (Genbank Accession:

NM_005929). They showed no homology to other human genes as shown using the

BLAST program (http://www.ncbi.nlm.nih.gov/blast). As a control, a scrambled, non- specific transgene into the vector (pS-scrambled; Ambion) was cloned.

Human SK-Mel-28 melanoma cells were obtained from the American Type Culture

Collection (Maryland, USA) and cultured as described previously [Richardson 2000].

Transfections of pS-MTf transgenes or pS-scrambled vectors were performed with

Lipofectamine™ 2000 reagent (Invitrogen) and cells were selected and maintained in

1000 μg/mL of G418 (Invitrogen). The work on this siRNA-mediated MTf down- regulation cell line was performed by Miss Louise Dunn (Iron Metabolism and

Chelation Program, Department of Pathology, University of Sydney, Australia) but is incorporated within this Chapter to clearly demonstrate and support the significant findings on the role of MTf in growth and proliferation.

3.2.8. RT-PCR and Western analysis

RNA was isolated using the TRIzol® reagent as described above and RT-PCR analyses of transcripts was carried out by standard procedures [Le & Richardson 2004] using the primers in Table 3.1. Briefly, 0.3 – 1.0 μg of RNA was incubated with gene-specific oligonucleotides (0.2 μM final primer concentration) in a 50 μL volume containing 25

™ μL of 2X Reaction Mix (1.6 mM MgSO4 and 200 μM dNTP) and 2 μL of SuperScript

III RT/ Platinum® Taq Mix for 30 min at 50°C. After reverse transcription the samples were initially denatured for 3 min at 94°C. The reactions were then amplified for 20 –

35 PCR cycles that included a 90°C denaturation step for 30 s, 55-60°C annealing step for 30 – 60 s, a 68°C extension step for 60 s, with a final extension time of 5 min. As an

internal control, β-actin was amplified from the same samples. Identities of transcripts were confirmed by DNA sequencing [Sekyere et al. 2005].

A peptide sequence at the C-terminal of mouse MTf (Ac-

DDHNKNGFQMFDSSKYHSQDLC-amide) was chosen based on antigenicity and probability of surface exposure. The conjugated peptide was used to immunise rabbits and the antibody was purified (Quality-Controlled Biochemicals, MA, USA). Protein isolation and Western analysis was performed as described in Section 2.6 [Gao &

Richardson 2001].

Table 3.1. Primers for amplification of mouse and human mRNA.

Oligonucleotides (5' - 3') Product Pair Primer Accession Size No. Name No. Priming Sequence (bp) Sites F- AGGCTCCTGAGCGTGACTT 121-139 1 mMTf NM_013900 693 R- CACTGTGCTGTGCTTCAGG 795-813 F- TCATCGCGGCCCAGGAGGCTG 266-286 2 hMTf NM_005929 892 R- GGCAGCCGGTTGGGGTCACAG 1137-1157 F- TCCCGAGGGTTATGTGGC 700-717 3 mTfR1 NM_011638 324 R- GGCGGAAACTGAGTATGATTGA 1002-1023 F- TCAGGTCAAAGACAGCGCTCA 833-853 4 hTfR1 NM_003234 488 R- TTGGGAATATGGAAGGAGACT 1300-1320 F- CCCGCCACCAGTTCGCCATGG 64-84 5 mβ-actin NM_007393 397 R- AAGGTCTCAAACATGATCTGGGTC 437-460 F- CCCGCCGCCAGCTCACCATGG 57-77 6 hβ-actin NM_001101 397 R- AAGGTCTCAAACATGATCTGGGTC 433-453 F- AGGTGGTAAAGGGTGGCTCC 203-222 7 hOAS1 NM_016816 70 R- ACAACCAGGTCAGCGTCAGAT 252-272 F- GCCCTGATGCTGACGATT 629-646 8 mMef2a NM_001033713 511 R- TTCCGACTGTTCATTCCAAG 1120-1139 F- TCTAGACATTGAGTCTCACTCTACCC 52-77 9 hMEF2A NM_005587 232 R- TTCTACAAGATCTGAGGTCCAGAG 260-283 F- GGCTAATGGTGGTTTCTG 1557-1574 10 mGls XM_129846 499 R- TTCAACATGACCCTCTGC 2038-2055 F- GGCTAATGGTGGTTTCTG 1532-1552 11 hGLS NM_014905 690 R- TTCTTATGGACGGTTTGA 2201-2221 F- GGGGCTCATACTCATCTT 843-860 12 mTcf4 NM_013685 787 R- GCCTGTCCTCCATTTCTA 1612-1629 F- GGCGGTTTGTGTGATTTTG 176-194 13 hTCF4 NM_003199 237 R- ATAGTTCCTGGACGGGCTTG 393-412 F- TTCTTTGCTTTGCGTTCC 747-764 14 mApod NM_007470 795 R- CGGTGGTCAGGGATTCTA 1524-1541

3.2.9. Labelling of transferrin with Fe

Established techniques were used to label apo-Tf (Sigma Aldrich, MO, USA) or apo-

MTf (from Malcolm Kennard, Synapse Technologies, British Columbia, Canada) with

56Fe or 59Fe (Perkin-Elmer, Boston, MA; [Food et al. 2002]).

3.2.10. Iron uptake assay

Standard techniques were used to determine cellular 59Fe uptake from 59Fe-Tf [Food et al. 2002; Kennard et al. 1995; Richardson & Baker 1992a]. Briefly, cells were incubated with 59Fe-Tf ([Tf] = 0.75 μM; [Fe] = 1.5 μM) for up to 240 min at 37ºC. The culture plates were then placed on a tray of ice and the monolayer washed four times with ice- cold PBS. The cells were subsequently incubated with the general protease, Pronase (1 mg/mL; Sigma), for 30 min at 4ºC. The cells were removed from the plates using a plastic spatula and centrifuged at 14,000 rpm for 1 min. This procedure resulted in the separation of Pronase-insensitive (internalised) 59Fe in the cell pellet, while the Pronase- sensitive (membrane-bound) 59Fe remained in the supernatant. Previous studies have demonstrated that this latter technique is valid for measuring internalised and membrane-bound 59Fe-Tf uptake by cells [Richardson & Baker 1992a; Richardson &

Baker 1990]. The supernatant and pellet were separated and placed in separate counting tubes. Radioactivity was measured on a γ-scintillation counter (Wallac, Compugamma,

Finland).

3.2.11. In vitro cell proliferation and migration assays

Cell counts were performed using Trypan blue staining. DNA synthesis was measured via incorporation of 3H-thymidine (Perkin-Elmer, MA, USA) using standard procedures

[Richardson et al. 1994]. Cell migration assays were performed over 18 h using

Transwell® filters in 24-well plates (Corning Inc, NY, USA) pre-coated with 0.15% gelatin [Demeule et al. 2003]. As a second measurement of cell migration and proliferation, wound scrape assays were also implemented [Gordon-Thomson et al.

2005; Sossey-Alaoui et al. 2005].

3.2.12. Tumour biology in nude mice

Male nude mice (BALB/c-nu) were subcutaneously injected at one site on the right flank with 1 x 106 SK-Mel-28 cells, MTf/B1 or scrambled control cells. Tumour growth was measured by standard procedures [Balsari et al. 2004].

3.2.13. Statistical analysis

Excluding the statistical analysis of the microarray results described above, all data were compared using Student’s t-test. Results were expressed as mean ± standard error of the mean (SEM) or mean ± standard deviation (SD). Data were considered statistically significant when p<0.05.

3.3. Results

3.3.1. Generation of MTf -/- mice

MTf -/- mice were generated by homologous recombination in ES cells. In this procedure, exons 2, 3 and 4 that encode the Fe-binding domain [Sekyere & Richardson

2000] and the intervening introns of the MTf gene were replaced with the neor gene. In addition, exon 1 was cloned in the reverse orientation (Figure 3.1). This targeting strategy disrupts the Fe-binding domain, the translation initiation codon and promoter.

Targeted ES cells were used to generate germ line–transmitting chimeric male mice that were mated with C57BL/6 mice to generate heterozygous offspring containing the neor gene. Mouse genotypes were confirmed by Southern analysis of tail DNA digested with

NheI and probed with a 3’ probe located outside the region of homology (Figure 3.1). In

MTf +/+ mice, a band of 12.6 kilobase (kb) corresponding to the 2 normal alleles was detected (Figure 3.2.A). Using MTf -/- animals, a 8.3-kb band representing the targeted allele was observed (Figure 3.2.A). The MTf +/- mice showed bands of 8.3 kb and 12.6 kb, corresponding to the targeted and wild-type alleles (Figure 3.2.A).

Figure 3.1. Targeted disruption of the mouse melanotransferrin (mMTf) gene. Targeting strategy used for deletion of mouse MTf exons 2, 3 and 4 and reversal of part (443 bp) of the MTf promoter region and the translation initiation codon. The genomic structure of the wild-type MTf allele is shown at the top depicting the promoter region, 16 exons and intervening introns. Primers used for identification of the wild-type allele are denoted P1 and P2. Fragments used for Southern analysis are denoted as the 5’ and 3’ probes. The targeting construct is shown below the wild-type allele. In this construct, the 1.4-kb BamHI fragment containing exon 1, intron 1 and part of the promoter region is reversed. After homologous recombination, the neomycin resistance (neor) targeted allele contains an additional NheI restriction enzyme site introduced with the neor gene cassette (Neo allele). In the Neo allele, exons 2, 3 and 4 have been replaced with the neor gene cassette flanked by lox-P sites and exon 1 is reversed. The neo cassette was deleted from the genome by mating of MTf +/- males to B6-deleter females to obtain the knockout (KO) allele. The P3 and P4 primers detect the targeted allele. Adapted from [Sekyere et al. 2006].

Figure 3.2. Confirmation of mMTf ablation in the mouse at the genomic DNA, mRNA and protein levels. (A) Genotype identification by Southern blot analysis from mice containing the neor cassette. (B) Genotype identification by Southern blot analysis from mice where there has been cre-mediated deletion of the neor cassette. (C) Genotype identification by PCR analysis of mice with the cre-mediated deletion of the neor cassette detects a wild-type (MTf +/+) fragment of 1.2 kb (first lane), a fragment of 0.7 kb in the targeted knockout (MTf -/-; second lane) and fragments of 0.7 and 1.2 kb in the (MTf +/-) mice (third lane). (D) RT-PCR of mMTf and mouse transferrin receptor 1 (Tfrc) mRNA transcripts in MTf +/+, MTf -/- and MTf +/- mice. (E) Western analysis of mMTf and Tfrc in testis and intestine from MTf +/+, MTf -/- and MTf +/- mice. Results are representative results in a typical experiment from 3 separate experiments. Adapted from [Sekyere et al. 2006].

The neor- and cre-negative heterozygous males and females were used to generate

MTf +/+ and MTf -/- littermates. Southern analysis confirmed the generation of homozygote mice (Figure 3.2.B). Genomic DNA from mouse tail was digested with

NheI and probed with a 3’ probe that detected a 12.6-kb fragment in the null allele and a

19.0-kb fragment in the targeted knockout (KO) allele (Figure 3.2.B). Two bands were detected in MTf +/- mice corresponding to the targeted and wild-type alleles (Figure

3.2.B). Routine genotyping by PCR yielded a 1.2-kb fragment (MTf +/+) and a 0.7-kb fragment (MTf -/-; Figure 3.2.D).

To confirm the knockout strategy did not result in a gene product, RT-PCR was performed (Figure 3.2.D). Using primers specific for exons 1 and 6 that amplify across the deleted region of MTf, no product was detected in tissues from MTf -/- mice (Figure

3.2.D). In contrast, a product was detected in MTf +/+ and MTf +/- mice (Figure 3.2.D).

MTf mRNA expression in heterozygous mice was reduced to 25% to 36% of that in

MTf +/+ littermates (Figure 3.2.D), indicating no compensatory up-regulation. As a measure of Fe status, expression of TfR1 (also known as Tfrc in mice; [Sekyere &

Richardson 2000]) was compared between the genotypes and was found to be the same

(Figure 3.2.D). Hence, ablation of MTf expression does not affect Fe pools that control

TfR1 expression [Sekyere & Richardson 2000].

Western analysis was performed to determine whether MTf protein expression was ablated in MTf -/- mice (Figure 3.2.E). MTf protein at the size of 82-kDa was detected in

MTf +/+ tissues that was absent from MTf -/- mice (Figure 3.2.E). These data confirm successful ablation of MTf. As found for TfR1 mRNA levels, examination of TfR1

protein expression showed no significant difference between the genotypes (Figure

3.2.E).

3.3.2. Phenotypic characterisation of MTf -/- mice

The MTf -/- mice were viable and fertile, showed no physical abnormalities and developed normally. The MTf -/- mice displayed no differences in reproduction, growth, development or histology compared with MTf +/+ littermates. Offspring from MTf +/- matings had a distribution of 26% MTf +/+, 48% MTf +/- and 26% MTf -/- (n = 312). This was consistent with a normal Mendelian inheritance ratio of 25% wild-types, 50% heterozygotes and 25% homozygotes. Litter sizes of MTf +/+ matings were comparable to the average litter sizes of MTf -/- matings. There were no significant differences in mouse body weight from 3-15 weeks of age (Figure 3.3.A and B). For example, at 12 weeks of age, the weights of male MTf +/+ mice (26.5 ± 0.4 g; mean ± SEM, n = 22) were not significantly different to male MTf -/- littermates (27.5 ± 0.6 g, n = 27; Figure

3.3.A). Similarly, the weights of female MTf +/+ mice (21.2 ± 0.4 g, n = 14) were not significantly different to female MTf -/- littermates (20.4 ± 0.5 g, n = 18; Figure 3.3.B).

Behavioural assessment using intruder and open-field studies [Crawley 1999; Crawley

2003; Miczek et al. 1994] showed no significant difference between the MTf +/+ and

MTf -/- genotypes (data not shown).

Figure 3.3. Examination of growth rates, organ to body weight ratios and brain Fe levels of MTf -/- compared with MTf +/+ littermates. Mice were maintained on a basal rodent chow diet, watered ad libitum and routinely weighed each week. Over fifteen weeks there were no significant differences in total body weight or growth rates between either male (A) or female (B) MTf -/- and MTf +/+ littermates (n = 14-27). The organ to total body weight ratios at autopsy of the brain, kidney, liver, spleen and heart were examined in male (C) and female (D) MTf -/- and MTf +/+ littermates (n = 5). No differences were observed in the ratios. (A-D) Results are expressed as mean ± SD. To examine brain Fe levels, mice were maintained on either basal rodent chow (0.02% Fe) or high Fe chow (2.00%) for four weeks. The levels of Fe in the brain were then determined by ICP-AES. There were no significant differences in brain Fe levels in MTf -/- mice maintained on a basal or high Fe diet compared with MTf +/+ littermates in both males (E) and females (F). Results (E-F) are expressed as mean ± SD (n = 6).

Examination of organ to total body weight ratios at autopsy indicated no gross changes between genotypes (Figure 3.3.C and D). Histological assessment of major tissues, bone marrow and blood smears showed no differences when sections from MTf -/- and MTf +/+ littermates were stained with haematoxylin and eosin, Giemsa, or Prussian-blue (data not shown).

A panel of haematological indices (Table 3.2) and selected serum biochemistry parameters (Table 3.3) were assessed with no significant differences being detected between MTf -/- and MTf +/+ littermates except in alkaline phosphatase (ALP) levels in male animals. However, this significant (p<0.05) decrease in ALP levels was within the normal range [Olfert et al. 1993]. Moreover, this difference in ALP levels was not observed when comparing female MTf -/- and MTf +/+ mice. All the haematological and biochemical parameters assessed were within physiological limits observed in mice

[Olfert et al. 1993; Taconic 2004]. A broader serum biochemistry and haematological screen, which included amylase, creatine kinase-MB, cholesterol, glucose, triglycerides,

+2 +2 -3 urea, Ca /Mg /PO4 , prothrombin, partial prothrombin time and differential cell counts (including neutrophils, basophils and eosinophils) also showed no significant differences between MTf -/- and MTf +/+ littermates (Table 3.3 and 3.4).

Male and female mice were assessed as separate groups due to the known differences in

Fe metabolism between the two sexes [Krijt et al. 2004; Sproule et al. 2001]. All of the above-mentioned parameters were also assessed as part of a longevity study when mice

were aged 18 months old, with no significant differences being observed (data not shown).

Table 3.2. Haematological indices in MTf -/- mice compared with their MTf +/+ littermates at 12 weeks of age.

Male Female Parameter Units MTf +/+ MTf -/- MTf +/+ MTf -/- (n = 19) (n = 21) (n = 17) (n = 25)

Red Blood Cell (RBC) 1012/L 8.6 ± 0.2 9.0 ± 0.1 9.2 ± 0.1 9.1 ± 0.1

White Blood Cell (WBC) 109/L 6.8 ± 0.7 6.3 ± 0.5 5.6 ± 0.4 5.5 ± 0.3

Haemoglobin (HGB) g/L 134 ± 3 138 ± 2 146 ± 4 132 ± 2

Haematocrit (HCT) 0.45 ± 0.01 0.47 ± 0.01 0.48 ± 0.01 0.47 ± 0.01

Mean Corpuscular Volume fL 52.2 ± 0.5 52.1 ± 0.4 52.5 ± 0.4 51.9 ± 0.4 (MCV)

Values expressed as mean ± SEM.

Table 3.3. Selected serum chemistry indices in MTf -/- mice compared with their MTf +/+ littermates at 12 weeks of age.

Male Female Parameters Units MTf +/+ MTf -/- MTf +/+ MTf -/- (n = 8-11) (n = 9-17) (n = 5-7) (n = 5)

Ca2+ mmol/L 3.04 ± 0.11 2.93 ± 0.06 3.31 ± 0.13 3.60 ± 0.01

Mg2+ mmol/L 1.35 ± 0.03 1.40 ± 0.09 1.22 ± 0.02 1.59 ± 0.19

PO4 mmol/L 3.21 ± 0.08 3.19 ± 0.14 3.24 ± 0.16 4.24 ± 0.37

Glucose mmol/L 14.8 ± 0.7 13.5 ± 0.7 12.8 ± 1.8 16.1 ± 0.7

Creatinine μmol/L 45.0 ± 1.4 42.1 ± 1.0 42.8 ± 0.9 48.5 ± 0.9

Urea mmol/L 10.9 ± 0.5 13.1 ± 2.6 9.2 ± 0.9 8.9 ± 0.5

Protein g/L 46.3± 0.6 43.5± 0.6 46.1 ± 0.7 43.8 ± 0.4

Albumin g/L 29.6± 0.4 27.5± 0.6 31.9 ± 0.5 30.7 ± 0.4

Globulins g/L 16.7± 0.2 16.0± 0.4 14.4 ± 0.6 13.2 ± 0.3

Aspartate Aminotransferase U/L 110 ± 16 108 ± 9 116 ± 19 145 ± 33 (AST)

Alkaline Phosphatase (ALP) U/L 122 ± 7* 102 ± 6* 142 ± 10 160 ± 18

Alanine Aminotransferase U/L 38.0 ± 3.5 47.8± 8.3 27.0 ± 1.7 26.0 ± 3.0 (ALT)

Amylase U/L 2611 ± 159 2641 ± 280 2632 ± 421 2442 ± 178

Creatine Kinase (CK-MB) U/L 129 ± 15 115 ± 12 149 ± 51 180 ± 10

Cholesterol mmol/L 2.96 ± 0.21 2.92 ± 0.21 3.88 ± 0.50 2.55 ± 0.23

Triglycerides mmol/L 1.43 ± 0.34 1.21 ± 0.17 0.66 ± 0.01 0.68 ± 0.00

Values expressed as mean ± SEM.

* p<0.05 as determined by Student’s t-test.

Table 3.4. Differential cell counts in MTf -/- mice compared with their MTf +/+ littermates at 12 weeks of age.

Male Female Parameter MTf +/+ MTf -/- MTf +/+ MTf -/- (n = 9) (n = 7) (n = 8) (n = 6)

Polymorphonuclear Neutrophils (PMN) 33 ± 4 29 ± 4 18 ± 4 15 ± 3

Band Neutrophil 0 2 ± 0 0 0

Lymphocyte 63 ± 4 69 ± 4 81 ± 4 89 ± 95

Monocyte 2 ± 1 2 ± 0 2 ± 0 1 ± 0

Eosinophil 1 ± 0 2 ± 1 1 ± 0 1 ± 0

Basophil 0 1± 0 0 0

Nucleated Red Blood Cell 1 ± 0 0 0 1 ± 0

Values expressed as % total white blood cell count (mean ± SEM).

3.3.3. MTf expression is not essential for normal Fe metabolism

MTf is a Tf homologue and shares many of its characteristics, including the ability to bind Fe [Brown et al. 1982; Richardson 2000]. To determine whether MTf has a role in

Fe metabolism, serum and tissue Fe indices were measured in 12 week old MTf -/- and

MTf +/+ littermates. In male or female mice, serum Fe levels, Tf saturation, TIBC and tissue Fe levels were not significantly different (p > 0.05) in MTf -/- animals compared with MTf +/+ littermates (Tables 3.5 and 3.6).

Since MTf has an Fe-binding site that is almost identical to Tf [Baker et al. 1992;

Brown et al. 1982; Rose et al. 1986], it could bind other vital metals such as Cu2+ and

Zn2+ [Morgan 1981]. Hence, it was important to examine the effects of the MTf null allele on tissue Cu and Zn levels (Table 3.7 and 3.8). As found for Fe, no differences were observed comparing MTf -/- animals with their MTf +/+ littermates (Table 3.7 and

3.8).

Table 3.5. Serum Fe indices in MTf -/- mice compared with their MTf +/+ littermates at 12 weeks of age.

Male Female Parameters Units MTf +/+ MTf -/- MTf +/+ MTf -/- (n = 11) (n = 16) (n = 7) (n = 5)

Serum Fe μmol/L 58.5 ± 3.0 51.6 ± 3.4 61.2 ± 12.0 51.6 ± 4.8

Unsaturated Iron-Binding μmol/L 43.9 ± 1.8 41.7 ± 2.4 39.6 ± 1.2 43.7 ± 3.9 Capacity (UIBC)

Total Iron-Binding μmol/L 102.5 ± 4.3 96.1 ± 5.3 100.8 ± 12.4 98.4 ± 8.7 Capacity (TIBC)

Transferrin saturation % 57 ± 1 55 ± 1 59 ± 3 55 ± 1

Values expressed as mean ± SEM.

Table 3.6. Tissue Fe stores in MTf -/- mice compared with their MTf +/+ littermates at 12 weeks of age.

Male Female Tissue Fe level (μg/g) MTf +/+ MTf -/- MTf +/+ MTf -/- (n = 25) (n = 26) (n = 15) (n = 17)

Liver 276 ± 14 291± 17 368± 17 426± 20

Spleen 3241 ± 186 2864± 166 4858± 396 4202± 392

Brain 69 ± 4 66± 4 69± 4 67± 3

Heart 643 ± 32 588± 32 573± 31 534± 28

Kidney 259 ± 11 243± 12 331± 22 309± 22

Values expressed as mean ± SEM.

Table 3.7. Tissue copper (Cu) stores in MTf -/- mice compared with their MTf +/+ littermates at 12 weeks of age.

Male Female Tissue Cu level (μg/g) MTf +/+ MTf -/- MTf +/+ MTf -/- (n = 15) (n = 18) (n = 8) (n = 5)

Liver 13.2 ± 1.0 15.6± 1.0 9.6± 0.9 9.7± 0.8

Spleen 1.4 ± 1.0 2.0± 0.7 1.9± 0.9 3.5± 3.2

Brain 26.4 ± 2.1 17.5± 1.0 24.6± 2.4 20.1± 1.6

Heart 11.6 ± 1.3 11.0± 0.9 9.4± 1.7 6.1± 1.2

Kidney 15.8 ± 1.1 11.7± 0.3 12.9± 1.6 11.0± 0.8

Values expressed as mean ± SEM.

Table 3.8. Tissue zinc (Zn) stores in MTf -/- mice compared with their MTf +/+ littermates at 12 weeks of age.

Male Female Tissue Zn level (μg/g) MTf +/+ MTf -/- MTf +/+ MTf -/- (n = 15) (n = 18) (n = 8) (n = 5)

Liver 79.4 ± 2.1 83.3± 5.6 78.7± 2.8 78.2± 3.4

Spleen 63.4 ± 5.2 74.6± 1.9 70.0± 2.8 66.7± 3.5

Brain 53.1 ± 4.1 47.6± 5.7 57.3± 5.0 54.7± 8.4

Heart 55.1 ± 4.3 59.5± 2.1 52.5± 7.9 51.6± 3.2

Kidney 63.8 ± 1.2 57.9± 0.9 59.7± 2.2 56.3± 1.2

Values expressed as mean ± SEM.

Considering the lack of phenotypes, the effect of the MTf null allele on Fe homeostasis was assessed by dietary Fe challenge of MTf -/- and MTf +/+ littermates. Mice were maintained on a high-Fe (2.00% Fe wt/wt) or basal Fe diet (0.02% Fe wt/wt) for 4 weeks, after which serum Fe indices and tissue Fe were measured. Animals on a high-

Fe diet showed a significant increase (p<0.001) in serum Fe (data not shown). When tissue Fe was examined in mice on a high-Fe diet, there was significantly (p<0.001) increased Fe loading in the spleen and liver of MTf -/- and MTf +/+ genotypes compared with animals given the basal Fe diet (Table 3.9). However, there was no difference in the response to the high-Fe diet between MTf -/- and MTf +/+ mice, both genotypes becoming similarly Fe loaded (Table 3.9).

To address the suggestion that MTf may play a role in brain Fe transport [Beliveau

2003; Demeule et al. 2002; Moroo et al. 2003], brain Fe levels were assessed in mice maintained for 4 weeks on either a basal (0.02%) Fe or high (2.00%) Fe diet (Figure

3.3.E and F). Regardless of diet, no significant differences were observed in either male or female MTf -/- mice compared with MTf +/+ littermates. Hence, the MTf null allele did not have any effect on Fe homeostasis.

Table 3.9. Tissue-Fe stores in MTf -/- mice compared with their MTf +/+ littermates at 12 weeks of age on basal-Fe (0.02% w/w) and high-Fe (2.00% w/w) diets.

Tissue Fe level (μg/g) Fe diet and n genotype by sex Liver Spleen Heart Kidney

0.02%

Male

MTf +/+ 5 286 ± 14 2628 ± 223 454 ± 52 210 ± 7

MTf -/- 6 280 ± 34 3426 ± 398 481 ± 47 396 ± 114

Female

MTf +/+ 6 552 ± 66 5565 ± 915 449 ± 43 374 ± 102

MTf -/- 6 451 ± 24 4949 ± 544 512 ± 34 361 ± 58

2.00%

Male

MTf +/+ 5 4018 ± 168 5274 ± 424 590 ± 51 466 ± 160

MTf -/- 6 3599 ± 378 4838 ± 365 702 ± 108 594 ± 164

Female

MTf +/+ 6 4317 ± 184 8145 ± 554 683 ± 95 556 ± 129

MTf -/- 6 3872 ± 280 6701 ± 485 579 ± 35 468 ± 16

Values expressed as mean ± SEM.

3.3.4. Comparative data analysis of the differential gene expression between MTf +/+

and MTf -/-mice

Considering that there were no observable phenotypic differences between the MTf -/- and MTf +/+ littermates, a microarray analysis was embarked upon to determine if there were changes in gene expression which may lead to a further understanding of MTf function. The levels of gene expression were determined from brain tissue RNA samples which were hybridised to the Mouse Genome 430 2.0 array GeneChip®

(Affymetrix). Mouse brain had the highest MTf expression in both sexes [Sekyere et al.

2005] and was used to determine differential gene expression between MTf -/- and

MTf +/+ mice. The top thirty genes showing both the largest positive and negative fold changes are summarised in Table 3.10. A positive value of the log2-fold change in expression indicates up-regulation in MTf -/- samples relative to the wild-type animals, while a negative value indicates down-regulation. Analysis of the genes affected by ablation of MTf (Table 3.10 and Figure 3.4) using NetAffxTM software

(http://www.affymetrix.com/analysis/index.affx) indicated they have roles in gene transcription, differentiation and development, regulation of cellular metabolism, transport and cell adhesion.

Table 3.10. The 30 genes showing the most significant differential expression between MTf -/- and MTf +/+ genotypes.

Gene Ratio Affymetrix ID Gene Title Symbol (log2)

1421252_a_at myocyte enhancer factor 2A Mef2a 0.91

1433670_at epithelial membrane protein 2 Emp2 0.63

1433911_at expressed sequence AI317223 AI317223 0.53

1434657_at glutaminase Gls 0.41

1425623_a_at cystathionine beta-synthase Cbs 0.31

1419281_a_at zinc finger protein 259 Zfp259 0.26

1458201_at Transcription factor 4 Tcf4 0.23

1419595_a_at gamma-glutamyl hydrolase Ggh -0.36

1416042_s_at nuclear autoantigenic sperm protein (histone-binding) Nasp -0.39

1418308_at Hus1 homologue (S. pombe) Hus1 -0.40

1443838_x_at fatty acid desaturase 2 Fads2 -0.41

1436608_at hypothetical protein MGC47262 MGC47262 -0.42

1451062_a_at peroxin 2 Pex2 -0.44

1451690_a_at poliovirus receptor-related 4 Pvrl4 -0.44

1422821_s_at StAR-related lipid transfer (START) domain containing 5 Stard5 -0.45

1438056_x_at ATP-binding cassette, sub-family C (CFTR/MRP), member 5 Abcc5 -0.47

1440983_at Tetratricopeptide repeat domain 15 Ttc15 -0.47

1419039_at cytochrome P450, family 2, subfamily d, polypeptide 22 Cyp2d22 -0.47

1416304_at LPS-induced TN factor Litaf -0.49

1434976_x_at eukaryotic translation initiation factor 4E binding protein 1 Eif4ebp1 -0.51

1423760_at CD44 antigen Cd44 -0.54

1418826_at membrane-spanning 4-domains, subfamily A, member 6B Ms4a6b -0.58

1438133_a_at cysteine rich protein 61 Cyr61 -0.62

1416821_at expressed sequence 2 embryonic lethal Es2el -0.63

1458710_at Polymerase (RNA) II (DNA directed) polypeptide A Polr2a -0.64

1432427_at NADH dehydrogenase (ubiquinone) 1 beta subcomplex 4 Ndufb4 -0.68

1418934_at mab-21-like 2 (C. elegans) Mab21l2 -0.70

1429262_at Ras association (RalGDS/AF-6) domain family 6 Rassf6 -0.78

1444564_at Apolipoprotein D Apod -1.81

Similar to the mRNA and protein data from the MTf -/- animal studies [Sekyere et al.

2005], microarray analysis demonstrated no differential expression of TfR1 or any other molecules known to be involved in Fe metabolism between MTf -/- and MTf +/+ mice.

These data relating to TfR1 expression, again confirm our previous observations that deletion of MTf did not affect the Fe pools that regulate TfR1 expression [Sekyere et al.

2005]. In addition, gene array assessment of MTf expression showed that it was down- regulated in the MTf -/- mice compared with their MTf +/+ littermates, as expected.

To validate the expression of the thirty genes showing the largest changes in the array data (Figure 3.4.A), RT-PCR was used. Increased expression of the following genes was observed in MTf -/- mice when compared with their MTf +/+ littermates: myocyte enhancer factor 2A (Mef2a), glutaminase (Gls) and transcription factor 4 (Tcf4; Figure

3.4.A). Conversely, expression of the apolipoprotein D (Apod) and MTf genes were down-regulated in MTf -/- mice compared with MTf +/+ littermates (Figure 3.4.A).

Densitometric analysis of the RT-PCR data was used to quantitate the magnitude of change in gene expression. Expression of Mef2a mRNA was found to be four-fold

-/- +/+ higher (log2 ratio of +2.03) in MTf mice compared with MTf littermates, while Gls and Tcf4 mRNA expression were increased by two-fold (log2 ratio of +1.03) and three- fold (log2 ratio of +1.63), respectively (Figure 3.4.B). Examination of Apod mRNA expression indicated a five-fold down-regulation (log2 ratio of -2.39; Figure 3.4.B) in

MTf -/- mice compared with MTf +/+ littermates. In addition to the samples submitted for gene array analysis, the data was also confirmed independently on RNA samples from other MTf -/- and MTf +/+ mice.

Figure 3.4. Comparative data analysis and RT-PCR of genes showing differential expression between MTf -/- and MTf +/+ mice. RNA from MTf -/- and MTf +/+ littermates was subjected to gene array analysis using the Affymetrix Mouse Genome 430 2.0 array GeneChips®. (A) RT-PCR of some of the genes assessed by their log2 value to show differential expression from the microarray data. (B) Densitometric analysis of RT-PCR. The bar graph represents the change in expression of genes affected as a result of MTf ablation. Results in (A) and (B) were performed on both RNA samples sent for gene array and independent samples from other mice. The results are representative of at least 3 separate experiments performed. Adapted from [Dunn et al. 2006].

3.3.5. Down-regulation of MTf expression in SK-Mel-28 cells by PTGS

While there were some differential gene expression in the microarray study, no observable phenotype was detected in the MTf -/- mice compared with MTf +/+ littermates. Indeed, the role of MTf may only become apparent in melanoma cells where it is expressed at very high levels [Brown et al. 1981a; Woodbury et al. 1980]. To assess this, SK-Mel-28 cells were stably transfected with a pS-MTf vector which encoded each of the four transgenes that transcribe a 19-mer anti-MTf hairpin siRNA. As a control,

SK-Mel-28 cells were transfected with the same vector containing a scrambled, non- specific sequence. Of the four siRNAs specifically targeted to the human MTf gene, the vector containing the B transgene (pS-MTf/B1) was found to consistently give the most pronounced down-regulation of MTf expression and inhibition of cellular proliferation.

The other three transgenes resulted in less MTf down-regulation and not as marked reduction in proliferation (data not shown). Therefore, all further experiments used clones from the B transgene. The fact that all transgenes reduced MTf expression to some degree suggested that the effect observed was not due to possible off-target effects of the siRNA. The clones obtained from transfection with the pS-MTf/B1 vector demonstrated down-regulation of MTf mRNA and protein by >90% (Figure 3.5.A) and

>80% (Figure 3.5.B), respectively, compared with cells transfected with the scrambled transgene. In Figure 3.5.B, two MTf protein bands were visible on the Western blot at

79- and 82-kDa which are the result of post-translational modification of MTf

[Desrosiers et al. 2003].

Figure 3.5. Post-transcriptional gene silencing (PTGS) of MTf in melanoma cells leads to decreased MTf mRNA and protein levels. SK-Mel-28 melanoma cells were stably transfected with a vector that contained a cloned transgene which encodes a short 19-mer double-stranded hairpin RNA to target MTf mRNA. As a control, a second vector containing a non-specific (scrambled) hairpin siRNA was similarly transfected into cells. Clones were selected in G418 and are referred to as “MTf/B1” and “scrambled control”. (A) RT-PCR analysis demonstrates marked down-regulation of hMTf mRNA in the MTf/B1 cell line compared with scrambled control cells. The expression of hTfR1 and hOAS1 remain unchanged. The transcription factors hMEF2A and hTCF4 were up-regulated in MTf/B1 cells. (B) Western blot analysis demonstrated a greater than 80% decrease in MTf protein expression in the MTf/B1 cell line compared with scrambled control cells. Results in (A) and (B) are representative of a typical experiment from 3 separate experiments performed. Adapted from [Dunn et al. 2006].

Multiple transfection experiments with the pS-MTf/B1 vector six months after the initial study, showed similar down-regulation of MTf expression in SK-Mel-28 cells, indicating that this was not a clonal effect. Cells transfected with a scrambled, non- specific transgene displayed no reduction in hMTf mRNA or protein expression compared with non-transfected SK-Mel-28 parent cells (data not shown). To exclude the possibility that siRNA was eliciting an interferon-stimulated response which could affect cellular phenotypes, RT-PCR was performed to assess the expression of the interferon-target gene, 2',5'-oligoadenylate synthetase 1 (hOAS1; [Bridge et al. 2003]).

There were no differences in hOAS1 expression between the scrambled control cell line,

MTf/B1 (Figure 3.5.A), or SK-Mel-28 cells (data not shown). As another control, the effect of down-regulation of MTf on the expression of the Fe-regulated gene, hTfR1

[Lambert et al. 2005a], was assessed. There was no alteration in hTfR1 mRNA expression in the MTf/B1 cell line compared with the scrambled control (Figure 3.5.A).

The expression of the transcription factors, hMEF2A and hTCF4, were assessed in

MTf/B1 and scrambled control cells. As found for the MTf -/- mice, these transcription factors were up-regulated in the MTf/B1 cell line (Figure 3.5.A). However, there was no significant change in hAPOD expression, while hGLS was not detected (data not shown).

During culture of MTf/B1 cells it became apparent their growth was slower than cells transfected with the scrambled, non-specific transgene. To quantitate this, proliferation was assessed using viable cell counts with Trypan blue staining. Over four days,

MTf/B1 cells showed decreased growth when compared with cells transfected with the scrambled transgene (Figure 3.6.A) and this was significant on days 2 (p<0.05), 3 and 4

(p<0.001). In contrast, there was no growth inhibition in scrambled control cells when compared with the parent SK-Mel-28 cells (data not shown). DNA synthesis in MTf/B1 cells was also examined using 3H-thymidine incorporation and was found to be significantly decreased (p<0.001) on days 1-4 compared with scrambled control cells

(Figure 3.6.B). Fluorescent staining for cell viability (using acridine orange, propidium iodide and Hoechst 33258) indicated that the depressed proliferation was not due to apoptosis or necrosis (data not shown).

Figure 3.6. Post-transcriptional gene silencing (PTGS) of MTf in melanoma cells leads to a significant decrease in cellular proliferation and DNA synthesis. SK-Mel-28 melanoma cells were stably transfected with a vector that contained a cloned transgene which encodes a short 19-mer double-stranded hairpin RNA to target MTf mRNA. As a control, a second vector containing a non-specific (scrambled) hairpin siRNA was similarly transfected into cells. Clones were selected in G418 and are referred to as “MTf/B1” and “scrambled control”. (A) Cellular proliferation was assessed by viable cell counts using Trypan blue. MTf/B1 cells showed depressed growth when compared with cells transfected with the scrambled transgene. (B) 3H- thymidine incorporation was significantly decreased in the MTf/B1 cell line on days 1 to 4 compared with cells transfected with the scrambled control vector. Results in (A) and (B) are presented as mean ± SD (6 determinations) in a typical experiment from at least 3 separate experiments performed. Adapted from [Dunn et al. 2006].

3.3.6. Addition of Fe, apo-MTf or holo-MTf does not rescue depressed proliferation

in melanoma cells with decreased MTf expression

To further assess the hypothesis that MTf may potentially act as an Fe transport molecule to aid melanoma proliferation [Brown et al. 1982], the ability of exogenous Fe to restore the depressed growth of MTf/B1 cells was examined. To do this, cells were incubated with ferric ammonium citrate (FAC; 100 μg/mL) that is well known to effectively donate Fe to cells [Goto et al. 1983; Richardson & Baker 1992a]. Previously,

FAC was shown to donate Fe to the same cell type used in this study (SK-Mel-28 cells) and down-regulate both hTfR1 mRNA and protein levels [Richardson & Baker 1992a;

Richardson 2000]. These studies with FAC resulted in no restoration or increase of proliferation in MTf/B1 cells or cells transfected with the scrambled siRNA transgene

(Figure 3.7.A). As a positive control for the ability of FAC to donate Fe to cells, hTfR1 mRNA expression was assessed and shown to be markedly decreased after an incubation of 1-4 days with this Fe source (data not shown). The lack of effect of FAC on proliferation suggested the cells were Fe-replete probably because they were grown under optimal conditions in media containing 10% FCS. The inability of FAC to rescue the slower growth of MTf/B1 cells suggested that its depressed proliferation was not due to Fe-deprivation because of decreased MTf expression.

Figure 3.7. Addition of Fe, soluble apo-MTf or soluble holo-MTf does not rescue decreased melanoma cell proliferation induced by MTf down- regulation. SK-Mel-28 melanoma cells were stably transfected with control vector expressing scrambled non-specific control siRNA or the anti-MTf siRNA transgene to generate MTf/B1 and scrambled cells respectively. (A) Proliferation rates were measured using viable cell counts after addition of Fe as ferric ammonium citrate (FAC; 100 μg/mL). Exogenous Fe did not restore proliferation in cells showing down-regulation of MTf. (B) There was no significant difference in 59Fe uptake from 59Fe-transferrin (59Fe-Tf; 0.75 μM) as a function of time between the scrambled control cell line and the MTf/B1 cell type with low MTf expression. (C) Addition of holo-MTf (1.0 μM) or (D) apo-MTf (1.0 μM) did not rescue the depressed proliferation rate of the MTf/B1 cell line compared with the scrambled control. Results are mean ± SD (3 determinations) in a typical experiment from 3 separate experiments performed. Adapted from [Dunn et al. 2006].

As a measure of Fe status in cells cultured under standard growth conditions, hTfR1 mRNA expression was assessed and found not to be increased in the MTf/B1 cell line compared with cells transfected with the scrambled control vector (Figure 3.5.A). These results showed that the MTf/B1 cells with low MTf expression were not Fe-deprived compared with the scrambled control. When holo-Tf ([Fe] = 1.5 μM) was added as an

Fe-donor, it also did not stimulate growth (data not shown), again probably because these cells were Fe-replete.

To further examine the relationship between MTf function and Fe metabolism, 59Fe uptake assays was performed using 59Fe-Tf (0.75 μM). These studies showed no significant difference in 59Fe uptake from 59Fe-Tf between MTf/B1 cells and the scrambled control (Figure 3.7.B), suggesting MTf had a negligible role in Fe uptake from Tf. The fact that 59Fe-Tf uptake was not greater than the scrambled control agrees with the other data indicating no difference in hTfR1 mRNA expression between

MTf/B1 cells and the scrambled control (Figure 3.5.A). These results support the finding that MTf/B1 cells were not Fe-deplete due to low MTf expression.

A previous study suggested sMTf can inhibit SK-Mel-28 cellular migration and that the balance between the soluble and membrane-bound forms of MTf can regulate this process [Demeule et al. 2003]. Hence, it was important to determine if incubating these cells with sMTf could restore proliferation of MTf/B1 cells. However, addition of holo- sMTf (Figure 3.7.C) or apo-sMTf (Figure 3.7.D) at a concentrations well above those used in a previous study (1 μM; [Kim et al. 2001]) had little effect on MTf/B1 cells or

the scrambled control. This indicates that it is membrane-bound MTf which is important for proliferation.

3.3.7. Inhibition of MTf expression decreases cell migration

The suggestion that MTf may be involved in migration was assessed by examining the ability of melanoma cells to pass from the upper to the lower levels of a transwell filter

(Figure 3.8.A) and their ability to repair a “wound” in a monolayer (Figure 3.8.B). The

MTf/B1 cell line showed a significant decrease (p<0.01) in migration over 18 h compared with the scrambled control (Figure 3.8.A). Since the cells do not significantly proliferate over 18 h (Figures 3.6.A, 3.7.A,C-D), the effect was unlikely due to a difference in growth rate. The ability of the cells to proliferate and migrate was also assessed via a wound closure assay (Figure 3.8.B). This assay is an indirect measure of cell migration as the effects of proliferation must also be taken into account. Scrambled control cells had closed the wound in the monolayer after 60 h, whereas MTf/B1 cells only begun to close the wound after 72 h (Figure 3.8.B).

Figure 3.8. Down-regulation of MTf expression reduces SK-Mel-28 melanoma cell migration in vitro. (A) Cell migration rates through a transwell chamber over 18 h by SK-Mel-28 melanoma cells stably transfected with anti-MTf siRNA (MTf/B1) or scrambled control siRNA. (B) MTf/B1 cells showed reduced proliferation and migration assessed by the “wound” closure assay when compared with the scrambled control. Control cells closed the wound after 60 h, while MTf/B1 cells begin to close the wound after 72 h. The photograph shows a wound in the scrambled and MTf/B1 monolayers at the 60 h time point. Results in (A) and (B) are expressed as mean ± SEM (3 determinations) in a typical experiment from 3 separate experiments performed. Adapted from [Dunn et al. 2006].

3.3.8. Inhibition of MTf expression by PTGS results in depressed tumour growth in

nude mice

To examine the effect of decreased MTf expression on growth in vivo, the proliferation of MTf/B1 cells compared with SK-Mel-28 and scrambled control cells was assessed by subcutaneous injection into nude mice and the growth of the tumours examined

(Figure 3.9.A). After 30 days, the tumour volume arising from the MTf/B1 clone was significantly (p<0.04) smaller, namely 20% and 22% of that found for SK-Mel-28 or the scrambled control cells, respectively (Figure 3.9.A). In addition, using the MTf/B1 cell type, tumours only developed in four of ten animals, while nine of ten mice developed tumours when injected with SK-Mel-28 or the scrambled control cells.

Examination of MTf protein levels in the harvested tumours after thirty days demonstrated that MTf expression in the MTf/B1 cell type was 20% of that of the parent cell or the scrambled control cells (Figure 3.9.B). These results demonstrated that MTf expression remained depressed in the MTf/B1 cell line while in vivo.

Figure 3.9. Down-regulation of MTf expression decreases tumour growth in vivo. (A) Net tumour growth in nude mice injected subcutaneously with 1 x 106 SK-Mel-28 melanoma cells, scrambled control cells, or the MTf/B1 cell line. The MTf/B1 cell line with low MTf expression displayed the least net tumour growth after 30 days in vivo. (B) Expression of MTf protein in MTf/B1 cells remains decreased relative to parental SK-Mel-28 cells or scrambled control cell lines after 30 days of in vivo growth in nude mice. (A) Results are mean ± SEM (4-9 animals each determination) in a typical experiment from 2 separate experiments. (B) Results are a typical experiment from 2 performed. Adapted from [Dunn et al. 2006].

3.4. Discussion

There is no conclusive evidence regarding the function of MTf, despite its suggested role in diverse metabolic processes including Fe metabolism and cellular differentiation

[McNagny et al. 1996; Michaud-Levesque et al. 2005a; Moroo et al. 2003; Sala et al.

2002; Sekyere & Richardson 2000]. Since MTf possesses a functional Fe-binding site

[Baker et al. 1992; Brown et al. 1982], a role in Fe metabolism appears plausible, particularly in malignant melanoma cells that express high levels of this tumour antigen

[Brown et al. 1982]. In the current study, two models using ablation and down- regulation of MTf expression were developed to determine its function. First, a MTf -/- mouse model was generated by targeted disruption of the MTf gene to assess its function in vivo. Second, to examine the role of MTf in neoplastic cells, we down- regulated MTf expression by PTGS using SK-Mel-28 melanoma cells which express high MTf levels [Brown et al. 1981b].

These studies demonstrated that MTf -/- mice were viable, developed normally and showed no obvious morphological, histological, behavioural, haematological, or Fe status changes when compared with MTf +/+ littermates. In contrast to other studies indicating a role for MTf in brain Fe metabolism [Demeule et al. 2002; Kennard et al.

1996; Moroo et al. 2003], no obvious changes were observed in brain Fe levels between

MTf -/- and MTf +/+ animals (Figure 3.3.E and F). This supports the hypothesis that MTf had no essential role in normal Fe metabolism or Fe supply to the brain. In addition, other studies have not confirmed a role for MTf in brain Fe metabolism [Pan et al.

2004; Richardson & Morgan 2004].

Interestingly, disruption of another Tf homolog, namely Lf, also led to no phenotype

[Ward et al. 2003], suggesting these molecules do not play essential roles in Fe metabolism and may have other functions [Sekyere & Richardson 2000]. In contrast, loss of function of molecules involved in Fe metabolism [Fleming et al. 1998; Levy et al. 1999; Meyron-Holtz et al. 2004; Trenor et al. 2000; Wallace et al. 2005; Zhou et al.

1998] has dramatic consequences. Indeed, in the case of TfR1, there are lethal effects in utero [Levy et al. 1999]. Clearly, if MTf had an essential role in Fe homeostasis, significant alterations in Fe indices, if not a deleterious phenotype, would have been apparent.

The current investigation supports our laboratory previous in vitro studies using melanoma cells which demonstrated that membrane-bound MTf or sMTf did not play an important role in Fe uptake [Food et al. 2002; Richardson & Baker 1991a;

Richardson & Baker 1991b; Richardson 2000]. These results were confirmed by others using different cell types [Kriegerbeckova & Kovar 2000]. Furthermore, sMTf did not bind to TfR1 or TfR2 to donate Fe to cells [Food et al. 2002; Kawabata et al. 2004].

Other studies also suggest that MTf does not play a key role in Fe metabolism. For instance, unlike TfR1 which is up-regulated to supply Fe for DNA synthesis in the S phase [Neckers & Cossman 1983], MTf expression remained constant throughout the cell cycle [Kameyama et al. 1986]. Moreover, in contrast to TfR1, MTf is not regulated by cellular Fe status [Richardson 2000] and is not up-regulated in dividing cells

[Seligman et al. 1986]. Collectively, these data demonstrate that MTf is not essential for normal Fe homeostasis.

100

As there was no obvious phenotype in the MTf knockout mouse, gene array analysis was employed to identify any changes in the gene expression profile of mice with the

MTf null allele. These studies showed no changes in the expression of molecules known to be involved in Fe metabolism. However, the analysis did identify genes that were differentially expressed between MTf -/- and MTf +/+ littermates. The products of these genes have roles in processes such as transcription, differentiation and development, regulation of cellular metabolism, transport and cell adhesion. Subsequent analysis of these changes using RT-PCR confirmed that the genes Mef2a, Tcf4 and Gls were up- regulated, while the Apod gene was down-regulated in MTf -/- mice compared with their wild-type littermates.

Of these genes, two were identified as transcription factors, namely Mef2a and Tcf4

[Black & Olson 1998; Mann et al. 1999]. The Mef2a transcription factor is strongly expressed in muscle tissues and is involved in fetal cardiac development, calcium- signalling and the MAP kinase pathways [Black & Olson 1998; Naya et al. 2002; Zhao et al. 1999]. This transcription factor is suggested to be important in cell proliferation, differentiation and survival [McKinsey et al. 2002]. The Tcf4 gene product plays a role in the Wnt signalling pathway [Graham et al. 2001] and the regulation of melanocyte differentiation [Furumura et al. 2001]. Tcf4 is expressed at high levels in some cancers and functional genomic investigations demonstrated this transcription factor has a role in cell proliferation [Mann et al. 1999; Zhao et al. 2004]. Notably, genes such as solute carrier family 2a member 4 (Slc2a4) and Cd44 are directly regulated by Mef2a and

Tcf4 [Thai et al. 1998; Wielenga et al. 1999] and these array data showed that Slc2a4 and Cd44 were affected by MTf deletion. In fact, Cd44 was found in the top thirty genes

101

showing differential regulation after MTf deletion (Table 3.10). These observations suggest some functional links between MTf, Mef2a and Tcf4.

The other two genes showing differential expression, namely Gls and Apod, have also been demonstrated to have a role in proliferation and encode the proteins, glutaminase and apolipoprotein D, respectively. Glutaminase has been found to be highly expressed in the mitochondria of tumour cells immediately before the maximum rate of proliferation is achieved [Aledo et al. 1994; Kvamme et al. 2000; Medina et al. 1992].

Conversely, apolipoprotein D inhibits cellular proliferation and is associated with transporting sterols, steroids and arachidonic acid for tissue repair [Sugimoto et al.

1994; Terrisse et al. 1998]. The up-regulation of Gls and down-regulation of Apod in the knockout mouse compared with the wild-type indicates a direct or indirect response to gene deletion and suggests some role for MTf in proliferation. Under the current experimental conditions, no defect in growth or development was obvious and a phenotype may only be observed when the mouse is exposed to an appropriate stress.

Given these results from the MTf -/- mouse, it should be noted that this is a normal physiological system in a whole organism which has many existing compensatory mechanisms that may mask deletion of MTf expression. Potentially, in melanoma cells that express high levels of MTf, the function of the protein may become more evident when its expression is decreased. MTf is hyper-expressed in melanoma cells and much less so in other tumours [Brown et al. 1981a; Woodbury et al. 1980], suggesting a cell- type specific function. We used PTGS in human SK-Mel-28 cells and showed that MTf down-regulation impaired proliferation, DNA synthesis and migration in vitro and

102

decreased tumour growth in vivo in nude mice. This was not a clonal effect as transfection experiments performed 6 months after the initial study gave similar results.

Moreover, the decrease in MTf expression and melanoma proliferation was observed with four different siRNA constructs (see Section 3.2.7), indicating that these effects could not be explained by the off-target influence of one individual siRNA.

The addition of Fe donors such as FAC or holo-Tf did not rescue the impaired proliferation of MTf/B1 cells with low MTf expression. This indicated that decreased growth of cells with low MTf expression was independent of any effect on Fe metabolism and this was in agreement with the MTf -/- mouse studies and investigations in vitro [Richardson & Baker 1991a; Richardson & Baker 1991b; Richardson 2000;

Sekyere et al. 2005]. Further, there was no difference in Fe uptake from Tf between clones with low MTf levels and the scrambled control.

Considering the role of MTf in proliferation, it is notable that evidence from hyper- expression studies using melanoma cells suggested increased MTf expression results in accelerated growth [Estin et al. 1989]. In contrast, MTf hyper-expression in CHO cells did not increase proliferation [Kennard et al. 1995]. These data support the results reported in this Chapter suggesting the role of MTf may only be obvious in melanoma cells where it is highly expressed.

Potentially, the changes in gene expression observed in the MTf -/- mice compared with

MTf +/+ littermates could be involved in the altered proliferation observed in the PTGS experiments using melanoma cells. Assessment of gene expression by RT-PCR

103

indicated that in melanoma cells engineered with suppressed MTf expression, hMEF2A and hTCF4 were similarly up-regulated. These data support the relationship between

MTf and the transcription factors observed in the MTf -/- mouse. Despite the decrease of

Apod expression in MTf -/- mice compared with wild-type animals, there was no change in hAPOD expression in MTf/B1 cells compared with their scrambled control counterparts. Therefore, to rule out the possibility that decreased Apod expression in

MTf -/- mice may have resulted from events associated with MTf gene deletion, the expression of a gene at the same chromosomal locus as MTf and Apod was examined. Indeed, TfR1 is located on the same strand as MTf (714-kb apart) and the former gene showed no change in its expression when comparing MTf -/- and MTf +/+ mice [Sekyere et al. 2005]. Apod is located on the alternate strand as MTf (564-kb apart) and these latter results demonstrating no difference in TfR1 expression between the genotypes [Sekyere et al. 2005] suggest that the alteration in Apod expression was not due to a deletion in MTf. However, it cannot be ruled out that the deleted region of MTf itself may contain distal regulatory elements for nearby genes such as Apod that could affect its expression in the MTf -/- mouse model.

In summary, studies using MTf -/- mice conclusively demonstrate that MTf does not play an essential role in Fe metabolism or homeostasis. Differentially expressed genes such as Mef2a, Tcf4, Gls and Apod in the knockout mouse were identified. This suggested that MTf may play a role in growth and proliferation under normal physiological conditions. It was also shown using melanoma cells that MTf down-regulation resulted in inhibition of proliferation and cell migration in vitro and reduced tumour growth in

104

vivo in nude mice. These data indicate that MTf plays an important role in melanoma cell growth and tumourigenesis.

105