Human Genetics

, Volume 113, Issue 3, pp 202–210

Analysis of zinc transporter, hZnT4 (Slc30A4), gene expression in a mammary gland disorder leading to reduced zinc secretion into milk

Authors

  • Agnes Michalczyk
    • Centre for Cellular and Molecular Biology, School of Biological and Chemical SciencesDeakin University, Burwood Campus
  • George Varigos
    • Department of DermatologyRoyal Children's Hospital
  • Anthony Catto-Smith
    • Department of Gastroenterology and Clinical NutritionRoyal Children's Hospital
  • Rachael C. Blomeley
    • Centre for Cellular and Molecular Biology, School of Biological and Chemical SciencesDeakin University, Burwood Campus
    • Centre for Cellular and Molecular Biology, School of Biological and Chemical SciencesDeakin University, Burwood Campus
Original Investigation

DOI: 10.1007/s00439-003-0952-2

Cite this article as:
Michalczyk, A., Varigos, G., Catto-Smith, A. et al. Hum Genet (2003) 113: 202. doi:10.1007/s00439-003-0952-2

Abstract

Zinc deficiency, causing impaired growth and development, may have a nutritional or genetic basis. We investigated two cases of inherited zinc deficiency found in breast-fed neonates, caused by low levels of zinc in the maternal milk. This condition is different from acrodermatitis enteropathica but has similarities to the "lethal milk" mouse, where low levels of zinc in the milk of lactating dams leads to zinc deficiency in pups. The mouse disorder has been attributed to a defect in the ZnT4 gene. Little is known about the expression of the human orthologue, hZnT4 (Slc30A4). Sequence analysis of cDNA, real-time PCR and Western blot analysis of hZnT4, carried out on control cells and cells from unrelated mothers of two infants with zinc deficiency, showed no differences. The hZnT4 gene was highly expressed in mouthwash buccal cells compared with lymphoblasts and fibroblasts. The hZnT4 protein did not co-localise with intracellular free zinc pools, suggesting that hZnT4 is not involved in transport of zinc into vesicles destined for secretion into milk. This observation, combined with phenotypic differences between the "lethal milk" mouse and the human disorder, suggests that the "lethal milk" mouse is not the corresponding model for the human zinc deficiency condition.

Introduction

Zinc is an essential trace element required for normal growth and development. It is needed for the function of enzymes (Vallee and Auld 1990), nucleic acids (Berg 1986), lipid metabolism (Cunnane 1988), membrane integrity (Bettger et al. 1980) and in signalling in the brain (Frederickson et al. 2000; Xie and Smart 1991).

The early symptoms of zinc deficiency include dermatitis, diarrhoea, alopecia, loss of appetite and neuropsychiatric changes (Aggett and Harries 1979). Prolonged deficiency results in impaired growth, hypogonadism and impaired immune function (Prasad 1985).

One inherited form of zinc deficiency is acrodermatitis enteropathica, a rare, autosomal recessive condition with features similar to those of nutritional zinc deficiency. Acrodermatitis enteropathica was identified following the discovery that the symptoms could be alleviated by zinc supplementation (Moynahan 1974). Impaired intestinal absorption and reduced uptake of zinc by jejunal cells was found in patients with this disorder (Atherton et al. 1979; Weismann et al. 1979). Sequence analysis of DNA from patients with acrodermatitis enteropathica has implicated a gene, SLc39A4, in the pathogenesis of disease (Kury et al. 2002; Wang et al. 2002). The gene, designated hZIP4, a member of the ZIP family of genes is expressed in mouse enterocytes, where it may facilitate intestinal zinc uptake (Wang et al. 2002).

A form of zinc deficiency found in breast-fed babies is caused by reduced levels of zinc in maternal milk. This condition has been reported in pre-term babies (27–33 weeks' gestation) (Aggett et al. 1980; Connors et al. 1983; Glover and Atherton 1988; Heinen et al. 1995; Parker et al. 1982; Stevens and Lubitz 1998; Weymouth et al. 1982; Zimmerman et al. 1982) and less commonly in term babies (Bye et al. 1985; Khoshoo et al. 1992; Stevens and Lubitz 1998). Prematurity does not explain the zinc deficiency despite higher requirements of pre-term babies for zinc; however, prematurity may affect predisposition to zinc deficiency. Premature babies are in negative zinc balance at birth due to reduced capacity for gut absorption and increased zinc demand because of rapid growth (Dauncey et al. 1977; Vileisis et al. 1981; Widdowson et al. 1974). In the zinc-deficient neonates, analysis of maternal milk indicated zinc levels were less than 40% that of normal milk at matched weeks of lactation (Weymouth et al. 1982; Zimmerman et al. 1982). Oral zinc supplementation induced a remission of zinc deficiency in these babies. Maternal zinc deficiency was not responsible for low zinc levels in the milk and maternal zinc supplementation, in most cases, did not affect zinc levels in milk (Connors et al. 1983; Parker et al. 1982; Weymouth et al. 1982; Zimmerman et al. 1982) or maternal plasma zinc levels (Parker et al. 1982; Weymouth et al. 1982); see Dorea (2000) for a review.

The aetiology of the zinc deficiency of neonates fed on breast milk is distinct from acrodermatitis enteropathica in several ways. Zinc deficiency in the breast fed babies is caused by the low levels of zinc in the maternal milk. In acrodermatitis enteropathica, however, the maternal milk is protective and the symptoms of zinc deficiency develop after weaning (Aggett 1983). No impairment in zinc uptake in the gut was found in the breast-fed zinc-deficient babies (Aggett et al. 1980). This is in contrast to acrodermatitis enteropathica, where mucosal zinc uptake in the small intestine of patients was lower than normal (Atherton et al. 1979). Other zinc transport studies showed decreased uptake of zinc in cultured fibroblasts from patients with acrodermatitis enteropathica (Grider and Young 1996).

The reduced concentration of zinc in the breast milk of mothers with zinc-deficient babies suggests that a defect similar to that found in the "lethal milk" mouse mutant might be present (Piletz and Ganschow 1978). Pups nursed on "lethal milk" mutant dams developed dermatitis, alopecia and showed stunted growth, leading to death. A defect in the secretion of zinc from the mammary gland was demonstrated in these mice (Ackland and Mercer 1992; Lee et al. 1992) and a nonsense mutation in the ZnT4 zinc transporter was reported to be responsible (Huang and Gitschier 1997).

Pedigree analysis has indicated that the condition that predisposes mothers to produce zinc-deficient breast milk is inherited (Sharma et al. 1988). We considered that hZnT4, a member of the solute carrier family 30, denoted SLC30A4, a candidate for this disorder. This gene has been excluded as the cause of the acrodermatitis enteropathia disorder (Nakano et al. 2002). The expression of SLC30A4, referred to in this paper as hZnT4, has not previously been investigated in mothers of premature babies with zinc deficiency. To investigate whether mutations or changes in expression of hZnT4 were responsible for this condition, we carried out a study of two unrelated mothers with low zinc milk levels, whose babies had developed zinc deficiency.

Materials and methods

A skin biopsy from mother 1 was used to establish a culture of fibroblasts using the scratch method (Fowler 1984). Fibroblasts were cultured in Nunclon 75 cm2 culture flasks in BME supplemented with 10% fetal bovine serum (CSL, Melbourne, Australia). The cells were passaged when confluent using 0.05% trypsin-versene solution in phosphate buffered saline (PBS) (Sigma-Aldrich, Sydney, Australia). Transformed lymphoblasts were established from blood of mother 1 by transfection with SV40 large T antigen. Lymphoblasts were grown as suspended cultures in RPMI-1640 supplemented with 20% fetal bovine serum (CSL, Melbourne, Australia). The cells were pelleted when the cultures reached high density by centrifugation and distributed to new flasks. Buccal cells from mother 2 were collected from mouthwashes, where, after vigorous cheek brushing with toothbrush, buccal cells were obtained following rinsing of the mouth with deionised water. Cells were pelleted by centrifugation, rinsed three times with PBS and immediately frozen at −80 °C. Fibroblast, lymphoblast and buccal cultures from normal individuals cell were established as controls.

Total RNA isolation

Cultured fibroblast cells were trypsinised and pelleted by centrifugation. After two rinses in PBS, total RNA was prepared using the QIAGEN RNeasy Mini kit, following the manufacturer's instructions. Cultured lymphoblasts were pelleted by centrifugation, rinsed twice in PBS and processed according to QIAGEN RNeasy Mini kit manual. Frozen buccal cell pellets were resuspended in TE buffer [10 mM Tris-HCl (pH 8.0), 10 mM EDTA], 200 mM NaOH and 1% SDS were added to the sample and incubated for 5 min at RT. The solution was neutralised with 3.0 M potassium acetate (pH 5.5) and incubated for 5 min at 4 °C. The supernatant was collected by centrifugation at 10,000 g for 10 min and used for total RNA isolation as above. The RNA concentrations were estimated spectrophotometrically at wavelengths of 260/280 nm.

Reverse transcription

Five micrograms of total RNA from each sample was reverse transcribed using 40 U AMV reverse transcriptase, RT buffer, 0.08 A260-units of random hexamer primer, 1 mM dNTP and 20 U RNase inhibitor (Roche Diagnostic, Sydney, Australia). The reaction was carried out at 41 °C for 1.5 h.

PCR and sequencing

The forward and reverse primers (see Table 1), at 55 pmol, were added to the PCR mixture consisting of: 5 μl of cDNA, 200 μM of each dNTP, PCR buffer, 1.5 mM MgCl2 and 1 U Taq DNA polymerase (Sigma-Aldrich, Melbourne, Australia). The following PCR amplification condition were applied: one cycle at 94 °C for 3 min, 34 cycles at 94 °C for 30 s, annealing temperature for 30 s, 72 °C for 60 s, with final extension at 72 °C for 10 min. PCR products were sequenced automatically [ABI 373 system using Thermo Sequenase Dye Terminator mix v. 3 (Amersham Pharmacia Biotech, Sydney, Australia)].
Table 1.

Primers. The primers for ZnT4 were designed to GenBank sequence accession no. AF025409 and those for β-actin were designed to GenBank sequence accession no. E00829

Forward Primer

Reverse Primer

PCR product length (bp)

ZnT4-C GGAAGCGCCTCAAATCTATGCT

ZnT4-D TACATTCAAAATGGCTTGGCACA

1,004

ZnT4-E GATAGCCTGGCAGTGAGAGCTG

ZnT4-F ATACGGACACAGCTGTCAGGGA

610

ZnT4-J TCGCTCAGCCTCTGCCATG

ZnT4-L GGTTTGCTTTGGACTGTACTT

1,216

ZnT4-M CAGAGTGTTGGTGTGCTAATA

ZnT4-N TTGAGAGACTGATGCTTGA

574

ZnT4-RealF2 aaccagtctggtcaccgtca

ZnT4-RealR2 CTATCCTGCCCATGGTTACG

93

βAC-RealF GACAGGATGCAGAACGAGAT

βAC-RealR TGATCCACATCTGCTGGAAGGT

138

Real-time PCR

Amplification reactions were performed with 1× SYBR Green PCR Master Mix (Applied Biosystem, Warrington, UK), 5 μM Z4TRT-1 forward and Z4TRT-2 reverse primers sequences (see Table 1) and 20 ng cDNA (RNA content prior to RT step). Samples were analysed in triplicate in 20 μl total volume using GeneAmp 5700 Sequence Detection System (PE Biosystems, Foster City, Calif.). An internal control of β-actin (see Table 1) was used to normalise RNA quantities and efficiency of reverse transcription. Amplification was performed as follows: one cycle at 50 °C for 2 min and 95 °C for 10 min, 40 cycles at 95 °C for 15 s and 60 °C for 60 s. Fluorescence produced by incorporation of SYBR green dye into double stranded DNA was recorded after the elongation phase of each repetitive cycle. The specificity of each reaction was determined by analysis of the melting-point dissociation curve generated at the end of each PCR. The threshold cycle value (CT), defined as the cycle number when fluorescence level exceeded the threshold value, was calculated after each reaction. The CT value of β-actin was subtracted from the hZnT4 CT value to produce ΔCT for each sample. A t-test was applied for statistical analysis of the results.

Western blot

Cell pellets, prepared as described above, were lysed by sonication using 15 pulses at 8 W power output in 500 μl of 10 mM Tris-HCl buffer with 1% SDS (Sigma-Aldrich, Melbourne, Australia). Immediately after sonication, EDTA-free protease inhibitor cocktail (Roche Diagnostic, Sydney, Australia) and 5 mM β-mercaptoethanol (Sigma-Aldrich, Melbourne, Australia) were added to the extracts. Samples were centrifuged for 5 min at 14,000 g and the supernatants collected. Protein concentrations were estimated using Bio-Rad DC Protein assay and spectrophotometry at a wavelength of 595 nm. Sixty micrograms of each extract was fractionated by SDS-PAGE using a Bio-Rad Mini Protean Gel system according to manufacturer's instructions. Proteins were transferred to nitrocellulose membranes (Pal Gelman, Melbounre, Australia) at 10 V for 25 min using a Trans-Blot SD semi-dry Transfer Cell (Bio-Rad, Sydney, Australia) in 25 mM Tris, 200 mM glycine (pH 8.3), 20% methanol and blocked overnight with 1% casein in TBS (0.05 M Tris and 0.15 M NaCl). The hZnT4 CSK antibody, raised to synthetic peptide (Chiron Technologies, Melbourne, Australia) and affinity purified as described previously (Michalczyk et al. 2002), was added to membranes at 1:200 dilution and incubated for 2 h at RT. After washes in blocking buffer (4× 15 min), hZnT4 protein was detected using 1:1,000 dilution of horseradish peroxidase-conjugated sheep anti-rabbit antibody and chemiluminescence detection kit (Roche Diagnostic, Sydney, Australia) according to manufacture's instructions. The membrane was placed in contact with Kodak XO-mat X-ray film for 1–60 min. After the first exposure, the membranes were stripped with Re-Blot Plus Western Blot Recycling Kit according to manufacturer instruction (Chemicon International, Calif., USA). Membranes were re-probed with β-actin antibody (1:2,000) and processed as above.

Immunofluorescence

Fibroblasts grown on 10-mm diameter glass coverslips were rinsed three times in PBS and fixed in 4% paraformaldehyde for 10 min. They were then rinsed twice in PBS, then permeabilised with 0.1% TX-100 in PBS for 10 min and blocked with 3% BSA in PBS for 90 min. Lymphoblast and buccal cells were smeared onto a glass slide coated with 5% gelatine, left to dry for 5 min and processed as above. Primary antibodies for hZnT4 (CSK) were diluted in 1% BSA in PBS (1:50) and were applied to cells overnight at 4 °C. Control cells were incubated with a mixture of 1 μM hZnT4 antibody with 4 μM hZnT4 CSK peptides overnight. After three PBS washes, a secondary antibody AlexaFluor 488 goat anti-rabbit IgG (Molecular Probes, Ore., USA), 1:2,000 dilution in 1% BSA in PBS was applied for 2 h at room temperature. Cells were washed three times for 10 min in PBS and 100 μM Zinquin [(ethyl-[2-methyl-8-p-toluenesulfonamido-6-quinolyloxy]acetate), generously supplied by Dr. Peter Zalewski, Department of Medicine, University of Adelaide, Adelaide, Australia] solution was added for 30 min at 37 °C. After three PBS washes, a drop of Fluoroguard (BioRad, Sydney, Australia) was added to the sections and a coverslip applied. Epifluorescence was viewed with an Olympus BX50 microscope with a PlanApo 60× 1.4 oil objective. Confocal images were collected using an Optiscan F900e system in a 512×512 pixel format.

Case history

Infant number 1, NM, was born premature at 36 weeks and was breast-fed for three months, without fortified milk. A rash appeared at two months of age, comprising a well-demarcated and symmetrical dermatitis, which affected cheeks, hands, fingers, buttocks and perineum. The infant was admitted to hospital, diagnosed as infected dermatitis, but failed treatment with antibiotics. Dermatological consultation raised the possibility of zinc deficiency and this was confirmed by tests carried out at three months postpartum. The zinc level in the blood of infant 1 was 4.5 μmoles/l (reference range 10.3–18.1 μmoles/l). Breast milk zinc from the mother of infant 1 (mother 1) was 0.29 μg/ml, which was less than one quarter that of the normal zinc level (1.35 μg/ml) at the corresponding stage of lactation. A skin biopsy showed a spongiotic epidermis and necrosis with inflammation. Bowel examination and biopsies were normal. Treatment with zinc (50 mg/day) resulted in a dramatic improvement in the rash within three days.

Infant number 2, TT, was born premature at 37 weeks gestation and presented with a slowly developing dermatitis affecting the face and perioral skin with scalp scale. The symmetrical appearance of the well-demarcated dermatitis was seen on the hands and fingers as well as the buttocks. Hair was very sparse, there was concern with failure to thrive and the child was weak and lethargic. The mother had breast-fed the child from birth without supplements. At seven months of age, following unsuccessful treatment for infection, dermatological consultation was sought. The clinical diagnosis of zinc deficiency was made on the basis of a typical rash. The maternal milk zinc level (mother 2) at seven months was 0.2 μg/ml, which was considerably less than the normal zinc level (0.85 μg/ml). On commencing treatment with zinc (50 mg/day), the rash cleared in three days and hair started to grow fully. Subsequently, the child dramatically improved, with weight gain and alertness.

Ethical consent for this study was granted by Deakin University and by the Royal Childrens Hospital.

Results

Screening of patient cells for mutations in hZnT4

Sequencing of hZnT4 cDNA was carried out on fibroblasts and lymphoblasts from mother 1 and buccal cells from mother 2 and appropriate controls corresponding to test samples.

There were no differences in cDNA sequences between controls and test samples in fibroblasts, lymphoblasts and buccal cells. In cDNA from both patients and eight controls we identified three base differences from the published hZnT4 sequence [GenBank AF025409 (Huang and Gitschier 1997)], one of which resulted in a GAG to GAC change, leading to a conservative substitution of a glutamic acid to an aspartic acid. A GAC is also present in the same position in hZnT4 sequences found in the genomic contig NT_010194 and several ESTs, including Z66008, Z66007 and Z62237. This suggests that the GAG in the AF025409 sequence is an error or possibly a polymorphism. Additionally, in our sequence, we found a difference in 23 nucleotides from the 3′UTR region of hZnT4, relative to the GenBank AF025409 sequence (Fig. 1).
Fig. 1.

Fragments of hZnT4 cDNA sequence from mothers 1 and 2 with zinc deficient infants. Underlined nucleotides indicate the sites that were found to be different from the GenBank sequence (AF025409). A G→C modification at nucleotide position 88 converts glutamic acid to aspartic acid. Two other modifications of C→T at the positions 844 and 915 were 'silent'. In the 3′UTR region of the hZnT4 gene the 23 nucleotides were different from the GenBank sequence, starting 11 bp downstream of the TAA termination codon

Analysis of hZnT4 mRNA expression

Real-time PCR was used to identify differences in hZnT4 mRNA expression levels between test and control samples. No non-specific signal was detected in control reactions (Fig. 2A asterisk). The specificity of each reaction was determined by analysis of the melting-point dissociation curve at the conclusion of each PCR reaction. The dissociation curve analysis of the PCR product revealed a strong single peak of expected Tm. This indicated that the only fluorescence detected came from hZnT4 cDNA fragment (Fig. 2A).
Fig. 2A, B.

Real-time PT-PCR analysis of hZnT4 mRNA was performed using the SYBR Green detection system. A The dissociation curve analysis of the PCR product confirms the specificity of the hZnT4 primers (a strong single peak of the expected Tm), and indicates the absence of the foreign DNA contamination (flat line marked by an asterisk). The dissociation curve was constructed by slow ramping of the temperature from 60 to 95 °C while continuously collecting fluorescent data after the completion of PCR. The gradual melting of the PCR products is indicated by a decrease in SYBR Green dye fluorescence. The curve of the PCR products melting profile was obtained by plotting the first derivative of fluorescence versus temperature. B Comparison of the hZnT4 mRNA expression levels between mother 1 cultured fibroblasts and lymphoblasts and corresponding controls revealed no significant differences (P>0.05, t-test). No differences were observed in mRNA in buccal cells from mother 2 versus controls (P>0.05, t-test). The ΔCT was calculated by subtracting the β-actin cycle threshold (CT) value (the cycle number in which the fluorescence emitted exceeds the threshold level) from that of the ZnT4

Complementary DNAs from mother 1 fibroblasts and lymphoblasts were tested against cDNAs from five control fibroblasts and lymphoblasts. The CT (cycle threshold) value, defined as the cycle number in which the fluorescence emitted exceeds the threshold level, was calculated. To eliminate differences in cDNA loading, the CT of β-actin was determined and this value was subtracted from that of the ZnT4 to produce ΔCT for each reaction. No significant differences between mother 1 ZnT4 mRNA (fibroblasts=7.30±0.7, lymphoblasts=7.74±0.65) and mRNA from control samples (fibroblasts=6.98±1.08, lymphoblasts=7.07±0.87) were found (P>0.05, t-test) (Fig. 2B).

Similar results were obtained when ZnT4 mRNA from mother 2 buccal cells was tested against three corresponding controls. The ΔCT value at 3.31±0.88 (mother 2) and 3.75±0.85 (control) were not significantly different (P>0.05, t-test) (Fig. 2B).

Analysis of hZnT4 protein expression

Expression levels of hZnT4 protein in control and test cells were measured by Western blot analysis using the CSK (anti-hZnT4) antibody. A band of the predicted size of 47 kDa was detected in all cells tested (Fig. 3A). An additional weak band of size 72 kDa was seen in samples, when reducing agents were absent from sample buffer (data not shown). Similar levels of hZnT4 protein were found in fibroblasts and lymphoblasts from mother 1 and corresponding controls. The hZnT4 protein expression levels in buccal cells from mother 2 were similar to protein levels in control buccal cells. A housekeeping protein, β-actin was used to indicate the relative levels of protein loaded on the gel (Fig. 3B).
Fig. 3A, B.

Western blot analysis of proteins from two mothers with zinc deficient infants. A Expression levels of the hZnT4 protein were analysed using the CSK antibody (anti-hZnT4). A band of the predicted size of 47 kDa was detected in all cells tested. Similar levels of hZnT4 protein were found in fibroblasts (lane 2) and lymphoblasts (lane 4) from mother 1 and corresponding controls (lanes 1, 3). Expression of hZnT4 in buccal cells from mother 2 (lane 6) was also similar to control (lane 5). B An antibody to housekeeping human β-actin was use to indicate the relative levels of protein loaded on the gel from mothers 1 and 2 (lanes 2, 4, 6) and corresponding controls (lanes 1, 3, 5)

Intracellular localisation of hZnT4

A cytoplasmic, granular label with some perinuclear distribution was detected in all control cells (Fig. 4A, G, M), together with mother 1 fibroblasts (Fig. 4B), lymphoblasts (Fig. 4H) and mother 2 buccal cells (Fig. 4N) using hZnT4 antibody. There was no distinctive difference between the distribution of hZnT4 within test and control cells.
Fig. 4A, B.

Localisation of hZnT4 protein in the cells from two mothers of the zinc-deficient infants, in relation to intracellular pools of the labile zinc. The test and control cells were stained with the CSK (anti-hZnT4) antibody followed by 30-min incubation in 100 μM Zinquin. A cytoplasmic, granular label of the hZnT4 with some perinuclear distribution was detected in all control cells (A, G, M) together with mother 1 fibroblasts (B), lymphoblasts (H) and mother 2 buccal cells (N). There was no distinctive difference between the distribution of hZnT4 within test and control cells. The test and control fibroblasts (C, D) and lymphoblasts (I, J) labelled with Zinquin to detect labile intracellular zinc, showed a perinuclear distribution extending into the cytoplasm. Buccal cells from mother 2 (N) and control (M) demonstrated more punctate, cytoplasmic localisation of Zinquin. No obvious overlap between hZnT4 and Zinquin was detected in either of the cell lines (E, F, K, L, Q, R)

Both control and mother 2 buccal cells showed the granular, cytoplasmic localisation of Zinquin stain with some weak perinuclear label (Fig. 4O, P). In fibroblasts Zinquin had a more distinct perinuclear distribution, extending into the cytoplasm (Fig. 4C, D). Lymphoblasts were grown in suspended cultures and had considerable depth. In these small, round cells it was difficult to resolve the subcellular distribution of Zinquin label, however, it appeared to localise mainly around the nuclei (Fig. 4I, J).

No obvious overlap between hZnT4 protein and Zinquin was detected (Fig. 4E, F, K, L, Q, R). Despite similar punctate appearance of both hZnT4 and Zinquin in buccal cells, the granules mainly appeared in different locations within the cytoplasm (Fig. 4Q, R). Interestingly, in lymphoblast cells from test samples and control, the cells that showed a stronger label for Zinquin had less expression of hZnT4 and vice versa.

Discussion

The clinical presentations of two cases of zinc deficiency in premature breast-fed infants described here are similar to previous reports (Aggett et al. 1980; Connors et al. 1983; Glover and Atherton 1988; Heinen et al. 1995; Parker et al. 1982; Stevens and Lubitz 1998; Weymouth et al. 1982; Zimmerman et al. 1982). The low levels of zinc in the maternal milk and a response by the infants to zinc therapy indicate a defect in the maternal breast affecting zinc secretion into milk. Both cases illustrated the dramatic recovery of their condition on the addition of zinc to the diet and this was subsequently only required for about six months as zinc levels remained normal without additional oral zinc in both cases, confirming the diagnosis of low breast milk zinc levels as the cause.

The clinical picture of this disorder is similar to the "lethal milk" mouse. We, therefore, investigated the possibility that a defect in maternal hZnt4 was responsible for the production of zinc-deficient milk in the mothers of the two pre-term babies. Little is known about the expression of this gene in human cells. Sequence analysis of the reading frames of cDNA from lymphoblasts, fibroblasts and mouthwash buccal cells was carried out and this showed similar results between control individuals and two unrelated mothers of the infants with zinc deficiency. To screen for the effects of mutations in non-coding sequences that may influence mRNA expression (Wang et al. 2002) or mRNA stability (Guhaniyogi and Brewer 2001), we measured expression levels of hZnT4 mRNA using real-time PCR. No differences between hZnT4 mRNA levels in both mothers and controls were detected.

Western blot analysis was used to detect expression of hZnT4 protein in lymphoblasts, fibroblasts and buccal cells. These cells have not previously been shown to express hZnT4. The presence of hZnT4 protein in all three cell types suggests that hZnT4 is widely expressed in human cells and is consistent with previous studies showing expression of hZnT4 in kidney, gut (Kury et al. 2002) and breast (Michalczyk et al. 2002). Protein levels of hZnT4 were similar in extracts of lymphoblasts, fibroblasts and buccal cells from control individuals and from mothers of the infants with zinc deficiency. The bands were the same size, indicating that there were no posttranslational modification differences. A six-fold difference in the hZnT4 mRNA expression levels in buccal cells compared with lymphoblasts and fibroblasts was found. This pattern was also shown in the protein levels measured by Western blot, where buccal cells expressed four times the amount of hZnT4 protein compared with the lymphoblasts and fibroblasts. The higher levels of expression of hZnT4 in buccal cells may be related to their function as transport epithelia.

We considered the possibility that the hZnT4 protein function may be affected by interactions with other proteins, as hZnT4 has a leucine zipper. Previous data suggested hZnT4 may form a complex as a larger 72-kDa band, in addition to the 47-kDa band (Michalczyk et al. 2002; Murgia et al. 1999). If the 72-kDa band represents a complex of hZnT4 and another protein, a defect in the interacting molecule may affect hZnT4 activity and therefore explain the patient's phenotypes. In the current study, there was no difference in the size of the larger band in cell extracts from mothers relative to controls (data not shown), indicating that there were no gross mutations (insertions or deletions) in the interacting protein.

In lymphoblast, fibroblast and buccal cells, the intracellular localisation of hZnT4 protein was found to be cytoplasmic with a perinuclear component. This is consistent with our previous observations in breast epithelial cells (Michalczyk et al. 2002). The localisation of a protein may be affected as a result of a mutation; for example, in Wilson and Menkes disease, where the defective ATPases were unable to traffic and showed a different localisation from the normal ATPases (Ambrosini and Mercer 1999; La-Fontaine et al. 2001). The localisation of Znt4 protein was similar in patient and control lymphoblast, fibroblast and buccal cells.

We used the fluorophore Zinquin to establish the presence of pools of labile zinc, as these pools have been reported in pancreatic islet (Zalewski et al. 1994) and respiratory airway cells (Truong-Tran et al. 2000). We found that ZnT4 protein did not co-localise with Zinquin, suggesting that hZnT4 was not involved in the transport of zinc into intracellular pools of labile zinc (zincosomes), which may be in the milk secretory pathway. We have previously found that hZnT1 and hZnT3 are expressed in human breast cells (Michalczyk et al. 2002), and if the protein products of these genes were investigated and found in zinc-containing vesicles, hZnT1 and hZnT3 may be candidates for the zinc deficiency disease.

Our results suggest that the "lethal milk" mouse is unlikely to be the corresponding model for the human mammary zinc secretion disorder. There are some differences between the mouse and human disorders that support this conclusion. In old age, the mouse shows symptoms of zinc deficiency (Piletz and Ganschow 1978), while zinc deficiency in women with defective zinc mammary secretion has not been reported. Maternal zinc supplementation in the "lethal milk" mouse is effective in alleviating pup zinc deficiency (Piletz and Ganschow 1978), but maternal zinc supplementation in humans does not increase milk zinc levels (Connors et al. 1983; Parker et al. 1982; Weymouth et al. 1982; Zimmerman et al. 1982). Finally, utricular otoconia are absent in the "lethal milk" mouse (Erway and Grider 1984), but abnormalities in balance that might be a consequence of defective utricular otoconia have not been reported in humans.

In summary, we described two cases of zinc deficiency in neonates, caused by a maternal defect in the transport of zinc into milk. This disorder has similarities to the "lethal milk" mouse condition. We investigated the hZnT4 gene, as a possible candidate responsible for the human zinc deficiency disorder. We did not detect mutations or changes in expression in the hZnT4 gene, which could account for a defect in zinc secretion in the human mammary gland. Interestingly, the hZnT4 protein did not co-localise with intracellular pools of zinc detected with Zinquin, which may be in the vesicular secretory pathway. This suggests that the hZnT4 transporter may not be pumping zinc into zinc-containing vesicles that are destined for secretion. It is likely, therefore, that the maternal mammary gland defect in zinc secretion is different from the "lethal milk" mouse and that other candidate hZnT transporters may mediate zinc secretion from the mammary gland and mutations in one of these may underlie the mammary zinc transport defect.

Acknowledgements

We thank Dr Sharon La Fontaine for her very constructive comments during the preparation of the manuscript. This study was supported by funding from the Australian Research Council to M.L.A..

Copyright information

© Springer-Verlag 2003