Characterization of Recombinant Human Lactoferrin Secreted in Milk of Transgenic Mice*

(Received for publication, September 27, 1996, and in revised form, January 4, 1997)

Jan H. Nuijens Dagger §, Patrick H. C. van Berkel , Marlieke E. J. Geerts Dagger , Peter Paul Hartevelt Dagger , Herman A. de Boer , Harry A. van Veen Dagger and Frank R. Pieper Dagger

From Dagger  Pharming BV, Niels Bohrweg 11-13, and  Leiden Institute of Chemistry, Medical Biotechnology Department, Leiden University, 2333 CA Leiden, The Netherlands

ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES


ABSTRACT

Human lactoferrin (hLF) is an iron-binding protein involved in host defense against infection and severe inflammation. Transgenic mice were produced harboring either hLF cDNA or genomic hLF sequences fused to regulatory elements of the bovine alpha S1 casein gene. Recombinant hLF expressed in the milk of transgenic mice (transgenic hLF) was compared with natural (human milk-derived) hLF. Immunological identity of the two forms was shown by double antibody immunoassays and the absence of an anti-hLF antibody response in transgenic mice on hyperimmunization with natural hLF. Mono S cation-exchange chromatography and N-terminal protein sequencing of transgenic and natural hLF revealed identical cationicity and N-terminal sequences. SDS-polyacrylamide gel electrophoresis and absorbance measurements of purified transgenic hLF showed this protein was 90% saturated with iron, whereas natural hLF is only 3% saturated. The pH-mediated release of iron from transgenic hLF was not different from that of iron-saturated natural hLF. Unsaturated transgenic hLF could be completely resaturated upon addition of iron. Slight differences in mobility between transgenic and natural hLF on SDS-polyacrylamide gel electrophoresis were abolished by enzymatic deglycosylation. Binding of transgenic and natural hLF to a range of ligands, including bacterial lipopolysaccharide, heparin, single-stranded DNA, Cibacron blue FG 3A, and lectins, was not different.

Based on these observations, we anticipate that (unsaturated) rhLF and natural hLF will exert similar, if not identical, antibacterial and anti-inflammatory activity in vivo.


INTRODUCTION

Human lactoferrin (hLF)1 is a single-chain metal binding glycoprotein of Mr 77,000 belonging to the transferrin family (1, 2). Lactoferrin consists of two highly homologous lobes, designated the N- and C-lobe, each of which can bind a single ferric ion concomitantly with one bicarbonate anion (2). The amino acid sequence of hLF has been determined by protein and DNA sequencing (1, 3) and its tertiary structure elucidated by x-ray crystallography (2). N-Linked glycosylation of hLF is of the sialyl N-acetyllactosaminic type (4). Human LF was first isolated from milk (5) but has also been found in many other external secretions such as tears, saliva, pancreatic juice, vaginal secretion, and seminal fluid, as well as in the specific granules of neutrophils (6). Extensive in vitro and some in vivo evidence indicates that hLF is involved in host defense against infection and severe inflammation, most notably at mucosal surfaces such as those of the gastrointestinal tract (6). In vitro antimicrobial activities of LF include bacteriostatic activity (7), probably residing in the capacity for high affinity binding of iron; bactericidal activity (8) probably mediated through direct binding of LF to bacterial cell wall components such as lipopolysaccharide (LPS) (9, 10); synergy with other antibacterial milk components such as human lysozyme (10) and immunoglobulin (11); and fungicidal activity (12). Human LF has been shown to promote the growth of Bifidobacterium species (13), the predominant bacteria of the intestinal flora of healthy breast-fed infants. The in vitro anti-inflammatory activities of LF include inhibition of the formation of hydroxyl radicals by scavenging free iron (14), inhibition of cytokine production (15), and binding to the inflammatory mediator LPS (9, 10, 16). Lactoferrin has been shown to promote the growth of intestinal cells in vitro (17). The function of specific LF receptors on intestinal cells and lymphocytes is as yet unclear (6).

The prospect of a wide variety of potential applications in human healthcare has motivated several investigators to study the feasibility of large-scale production of functional recombinant hLF. Here we describe the characterization of hLF secreted in milk of transgenic mice. Transgenic hLF was found to be identical with natural hLF derived from human milk by immunological and functional criteria. Structurally it differed only in glycosylation and the degree of iron saturation.


MATERIALS AND METHODS

Reagents

CNBr-activated Sepharose 4B, S-Sepharose, concanavalin A-Sepharose, heparin-Sepharose CL-6B, chelating Sepharose fast flow and blue Sepharose CL-6B were obtained from Pharmacia Biotech Inc. (Uppsala, Sweden). Calf thymus single-stranded DNA-agarose was purchased from IBF Biotechnics (Villeneuve-la-Garenne, France). Ulex europaeus agglutinin (UEA) type I-agarose was purchased from Sigma.

Generation of Transgenic Mice

Mammary gland-specific expression vectors based on regulatory elements from the bovine alpha S1 casein gene and either hLF cDNA (3) or the genomic hLF sequences were introduced into the murine germline as described (18). Signal sequences were derived from the bovine alpha S1 casein gene (in case of cDNA-based vectors) or the hLF gene (genomic vectors). Comparison of the coding sequence of genomic hLF and published hLF sequences (19-22), revealed that the amino acid sequence of mature genomic transgenic hLF represents the most frequently occurring sequence for hLF.2 The cDNA encoding transgenic hLF contains two unique differences, Ile130 right-arrow Thr and Gly404 right-arrow Cys (3). A detailed description of the construction of hLF expression vectors, as well as of the differences in hLF milk concentrations between distinct categories of hLF transgenics, is beyond the scope of this paper and will be published separately.3 Milk from lactating females was obtained as described (18).

Radioimmunoassay (RIA) for hLF

A quantitative RIA for hLF with anti-hLF monoclonal antibody (mAb) 13.17 coupled to Sepharose was performed essentially as described (23). Briefly, mAb 13.17-Sepharose suspension was incubated for 16 h by head-over-head rotation with test samples. A minor modification was that serial dilutions of natural hLF (standard) and mouse milk were made in phosphate-buffered saline, 0.1% Tween 20, 5 M NaCl to disrupt electrostatic interactions of hLF with murine milk components. Sepharose beads were then washed and hLF bound to Sepharose was detected by incubation with polyclonal rabbit 125I-anti-hLF for 6 h. After washing, Sepharose-bound radioactivity was measured. Results were expressed as percentage binding of the total amount of labeled antibodies added.

Antibody Responses in Nontransgenic and hLF Transgenic Mice after Immunization with Natural hLF

Male and female nontransgenic mice, cDNA hLF (line 33) and genomic hLF transgenic mice (line 937) were given repeated intraperitoneal injections with 50 µg of purified natural hLF, either in complete Freund's adjuvant (first injection) or incomplete Freund's adjuvant (booster injections: 3, 6, and 9 weeks after the first injection). Essentially the same hyperimmunization protocol has previously been used to prepare murine mAbs (24). Ten days after the fourth injection, blood samples were obtained, and sera were tested for the presence of specific antibody by a RIA, in which goat-anti-mouse immunoglobulin coupled to Sepharose was incubated with serial dilutions of mouse sera together with 125I-hLF. The anti-hLF antibody responses of individual mice were expressed as percentage of the mean response, arbitrarily defined as 100%, in the nontransgenic mice.

Saturation of hLF with Iron

Freshly prepared iron-nitrilotriacetic acid solution was used to saturate hLF purified from fresh human milk (designated as "natural hLF") with iron as described (16). Saturation of natural hLF with iron (designated as "iron-saturated natural hLF") was complete as assessed by absorbance measurement and nonreduced SDS-PAGE of nonboiled samples (25).

SDS-PAGE

Nonreduced SDS-PAGE was performed as described (16). In some cases, samples were boiled for 5 min in nonreducing SDS sample buffer (26) prior to SDS-PAGE analysis to achieve concomitant denaturation and desaturation of hLF (25). Reduced SDS-PAGE was performed after boiling of samples for 5 min in SDS sample buffer containing 5% (v/v) beta -mercaptoethanol.

Absorbance Measurement and Absorbance Spectrum Analysis

Absorption coefficients of A0.1%280 = 1.1 and A0.1%280 = 1.4 were used for unsaturated and iron-saturated hLF, respectively, and A0.1%465 = 0.058 for iron-saturated hLF (27).

The concentration of iron-saturated hLF (in mg/ml) in purified hLF was calculated by dividing the A465 by 0.058. The concentration of unsaturated hLF (in mg/ml) was calculated as: (A280 - 1.4 × (A465/0.058))/1.1. The percentage of iron saturation of purified hLF preparations was calculated as the ratio of iron-saturated hLF to total hLF. The UV-VIS absorbance spectra of purified hLF preparations were recorded using an Ultrospec II spectrophotometer (LKB, Bromma, Sweden)

Release of Iron (Desaturation) from Iron-saturated hLF

Iron-saturated natural hLF, cDNA transgenic hLF, and genomic transgenic hLF were diluted 7-fold in buffers of varying pH to final concentrations of 4.2, 3.9 and 4.3 mg/ml, respectively. Samples were incubated at 20 °C and the A465 measured continuously. Results are expressed as percentage desaturation by reference to iron-saturated natural or transgenic hLF diluted in 0.15 M NaCl. Buffers were 0.1 M MES, 0.15 M NaCl of pH 6.1, 0.1 M sodium acetate, 0.15 M NaCl of pH 5.1, pH 4.5, pH 4.0, and of pH 3.7, and 0.1 M glycine-HCl, 0.15 M NaCl of pH 3.1, pH 2.5, and of pH 2.1. Complete removal of iron from transgenic and iron-saturated natural hLF was performed by dialysis of preparations against 0.1 M glycine-HCl, 0.15 M NaCl, pH 2.5 for 2 h at 30 °C.

RIA Procedures to Compare the Binding of Distinct hLF Species to Solid-phase Ligands

Serial dilutions of purified hLF (1 µg/ml) in phosphate-buffered saline, 0.1% Tween 20 were incubated with ligand coupled to agarose as described above for the RIA of hLF. Bound hLF was detected by subsequent incubation with polyclonal 125I-anti-hLF antibodies. The response with natural hLF in each RIA was arbitrarily defined as 100%. From the parallel dose-response curves, the reactivity of iron-saturated natural hLF, cDNA transgenic hLF, and genomic transgenic hLF to each ligand was expressed as a percentage by reference to natural hLF. All RIAs were performed with 0.5 ml of agarose suspension (2 mg of agarose/ml of phosphate-buffered saline, 0.1% Tween 20, 0.02% sodium azide). Ligand-agarose was either commercially obtained or prepared as follows: anti-hLF mAbs E1 and E2 were partially purified from hybridoma culture supernatant by precipitation with 50% ammonium sulfate and coupled to Sepharose (20 mg of protein to 1 g of CNBr-activated Sepharose 4B); 15 mg of purified rabbit anti-hLF was coupled to 1 g of Sepharose; Fe2+-Sepharose was prepared by incubating a solution of FeSO4 with chelating Sepharose fast flow; Salmonella minnesota Re595 LPS-Sepharose was prepared as described (16). The specificity of hLF interaction with solid-phase ligand was demonstrated by RIAs in which Sepharose to which no ligand had been coupled ("glycine-Sepharose") was used (16).


RESULTS

Immunological Identity of Transgenic and Natural hLF

Transgenic mice expressed hLF in a mammary gland-specific and lactation-restricted fashion (18) at levels in milk up to 13 mg/ml. This paper describes the characterization of the transgenically produced protein. Fig. 1 shows dose-response curves of natural hLF and milk samples of genomic hLF (A) and cDNA hLF (B) transgenic mice in the RIA for hLF. The identical slopes and maximal responses in the RIA indicate that the antigenic determinants for monoclonal and polyclonal anti-hLF antibodies in natural and transgenic hLF are equally well accessible. Natural and transgenic hLF thus appear immunologically identical, and levels of hLF in transgenic mouse milk can be calculated by reference to natural hLF.


Fig. 1. Dose-response curves of natural hLF and transgenic hLF in the RIA for hLF. Serial dilutions of natural hLF (10 µg/ml; A and B, bullet ), of nontransgenic mouse milk (B, ×), of genomic hLF transgenic mouse milk (A, black-square, square , and open circle ), and of cDNA hLF transgenic mouse milk (B, black-square, square , and open circle ) were incubated with anti-hLF mAb 13.17 coupled to Sepharose. Bound hLF was detected by subsequent incubation with polyclonal 125I-anti-hLF antibodies. Results are expressed as percentage binding of the 125I-anti-hLF antibodies added. The volume of the experimental sample tested (µl) is indicated on the abscissa. Human LF concentrations in the genomic hLF transgenic mouse milk samples were: 3.8 mg/ml (black-square, line 937), 0.3 mg/ml (square , line 984), 1.3 mg/ml (open circle , line 984). The hLF concentrations in the cDNA hLF transgenic mouse milk samples were: 0.8 mg/ml (square , line 649), 1.7 mg/ml (black-square, line 649), 0.4 mg/ml (open circle , line 660).
[View Larger Version of this Image (18K GIF file)]


The immunological identity of transgenic hLF with natural hLF was confirmed by immunizing of transgenic mice with natural hLF (Fig. 2). No anti-hLF response was detectable (lower than 0.25% of that in controls) in 9 out of 11 transgenic mice, whereas a strong antibody response was observed in all six nontransgenic mice (40,000- fold diluted sera still showed anti-hLF activity). These results indicate that natural hLF is considered "self" by the immune system in the transgenic mice.


Fig. 2. Antibody response in nontransgenic and hLF transgenic mice after repeated injections with natural hLF. Nontransgenic mice (n = 6), cDNA hLF (n = 6), and genomic hLF transgenic mice (n = 5) were repeatedly injected with purified natural hLF. The antibody responses in sera of individual mice (bullet ) were expressed as percentage of the mean response (arbitrarily defined as 100%) in the nontransgenic mice. The horizontal lines indicate the mean response in each group of mice.
[View Larger Version of this Image (15K GIF file)]


Purification and N-terminal Sequence Analysis of Transgenic and Natural hLF

Analytical chromatography of transgenic mouse whey and human whey on a Mono S cation-exchange column (16) revealed that transgenic and natural hLF had completely bound to the column and that all mouse and human whey proteins other than hLF had eluted at a salt concentration of 0.3 M NaCl. Therefore, transgenic hLF as well as natural hLF were isolated by batchwise extraction with S-Sepharose from milk to which 0.4 M NaCl had been added (16). Unless stated otherwise, "cDNA transgenic hLF" and "genomic transgenic hLF" refer to rhLF isolated from pooled milk of transgenic mouse lines 33 and 937, expressing hLF in milk at about 0.2 and 2.5 mg/ml, respectively.

Purified transgenic and natural hLF were then subjected to Mono S chromatography to exclude that N-terminal degradation of hLF had occurred during the purification process (16). Both cDNA (Fig. 3) and genomic transgenic hLF (not shown) as well as natural hLF (see Ref. 16) eluted as single peaks at exactly the same position (0.7 M NaCl), i.e. the position characteristic of uncleaved hLF as was confirmed by N-terminal protein sequence analysis. This analysis also revealed that the bovine alpha S1 casein signal sequence (cDNA hLF) and the hLF signal sequence (genomic hLF) had been correctly and completely removed in the murine mammary gland to yield mature rhLF.


Fig. 3. Mono S chromatography and N-terminal protein sequencing of transgenic hLF. Fifty µg of S-Sepharose-purified cDNA transgenic hLF was applied to a Mono S HR 5/5 (Pharmacia) column, and bound protein was eluted with a linear salt gradient from 0-1.0 M NaCl as described (16). N-terminal protein sequencing (16) results are provided by the standard one-letter code for amino acids.
[View Larger Version of this Image (14K GIF file)]


Iron Binding Properties of Transgenic and Natural hLF

The binding of metal ions to hLF is accompanied by a change in its conformation (2). The SDS-PAGE migration of iron-saturated hLF, which has a more globular structure (2), is faster than that of unsaturated hLF, if the analysis is carried out under nonreducing conditions with freshly diluted, nonboiled samples (25). Boiling in the presence of SDS brings about denaturation of the iron-saturated molecule with release of iron (25). Boiled samples of purified genomic and cDNA transgenic hLF migrated as two doublets of protein bands on nonreduced SDS-PAGE (Fig. 4, lanes 1 and 2), whereas only two bands were observed with natural hLF (lanes 3 and 4). The ratio of the minor to the major band in natural hLF was about 1 to 9 (lanes 3 and 4, see also Ref. 16), whereas the ratios of the minor to the major doublet in genomic and cDNA transgenic hLF were about 1 to 9 and 4 to 6, respectively (lanes 1 and 2). Results with nonboiled samples suggested that most of genomic and cDNA transgenic hLF was saturated with iron (lanes 5 and 6). Quantitation of iron in purified transgenic hLF preparations (28) yielded results identical to calculations of the iron content based on absorbance measurement (see below). On reduced SDS-PAGE, the mobility of the two bands in natural hLF and the two doublets in transgenic hLF decreased to the same extent (lanes 9-12). The ratios of bands and doublets representing iron-saturated hLF in nonboiled samples of natural and transgenic hLF (lanes 5-8) were identical to the ratios found in respective hLF samples from which iron had been released upon boiling (lanes 1-4) or reduction (lanes 9-12). SDS-PAGE analysis of deglycosylated natural and transgenic hLF showed a single protein band with the same mobility for each hLF species, indicating that glycosylation heterogeneity accounts for the presence of the distinct protein bands in glycosylated natural and transgenic hLF (Fig. 5).


Fig. 4. Nonreduced and reduced SDS-PAGE (7.5%) analysis of purified hLF. Nonreduced SDS-PAGE (lanes 1-8) of purified hLF (5 µg) freshly diluted in nonreducing SDS sample buffer and applied either directly (lanes 5-8) or after boiling for 5 min (lanes 1-4). Reduced SDS-PAGE (lane 9-12) of samples that were boiled for 5 min in reducing SDS sample buffer. Gen, genomic transgenic hLF; cDNA, cDNA transgenic hLF; Fe, iron-saturated natural hLF; Nat, natural hLF. Proteins were visualized by staining with Coomassie Brilliant Blue. Left-hand numbers (Mr × 10-3) indicate the migration of the protein standards.
[View Larger Version of this Image (46K GIF file)]



Fig. 5. Reduced SDS-PAGE (10%) of deglycosylated natural and transgenic hLF. Human LF was deglycosylated with N-glycosidase F (26). Five µg of untreated (lanes 1-4) and deglycosylated (lanes 5-8) hLF species were loaded. For explanation of abbreviations, see Fig. 4.
[View Larger Version of this Image (64K GIF file)]


Absorbance spectrum analysis of cDNA transgenic hLF revealed an absorption peak at 465 nm (Fig. 6), indicating that the metal ion bound to transgenic hLF is indeed iron (27). Genomic transgenic hLF showed a similar absorbance spectrum (results not shown). The A280/A465 ratio of natural hLF was 573.1, whereas those of iron-saturated natural hLF, cDNA, and genomic transgenic hLF were 24.2, 27.3, and 25.8, respectively. Calculation of the degree of saturation of hLF with iron indicated that natural hLF was 3% saturated, whereas iron-saturated natural hLF, cDNA transgenic hLF, and genomic transgenic hLF were 100, 86, and 92% saturated, respectively. The A410/A465 ratios of iron-saturated natural hLF, cDNA, and genomic transgenic hLF were 0.80, 0.85, and 0.82, respectively, indicating that preparations were pure (22).


Fig. 6. Absorbance spectra of iron-saturated natural hLF and cDNA transgenic hLF between 200 and 700 nm. Trace 1, iron-saturated natural hLF (16.1 mg/ml); trace 2, cDNA transgenic hLF (10.9 mg/ml); trace 3, natural hLF (18.6 mg/ml).
[View Larger Version of this Image (19K GIF file)]


Iron Release from Transgenic hLF and Resaturation with Iron of Desaturated Transgenic hLF

Iron is released from iron-saturated hLF under acidic conditions. Iron release from hLF begins at about pH 4.0 and is complete at about pH 2.5 (29). To obtain information on the iron binding characteristics of transgenic hLF, we compared the pH-mediated iron release from this protein with that of iron-saturated natural hLF, monitoring the decrease in A465 upon incubation in buffers over the pH range 6.1-2.1. The results indicated that the release of iron under acidic conditions from transgenic hLF is not different from that of iron-saturated natural hLF (Fig. 7). The pH values at which 50% of hLF molecules were desaturated (after 24 h) were 3.6, 3.7, and 3.6 for iron-saturated natural hLF, cDNA, and genomic transgenic hLF, respectively.


Fig. 7. Desaturation of iron-saturated natural hLF, cDNA, and genomic transgenic hLF at varying pH. Purified hLF samples were diluted in buffers of varying pH, and the absorption at 465 nm was measured after incubation for 24 h at 20 °C. Results are expressed as percentage desaturation by reference to iron-saturated hLF diluted in 0.15 M NaCl. The experiment was performed on four occasions with iron-saturated natural hLF (bars represent the mean and S.D.) and on one occasion with cDNA and genomic transgenic hLF.
[View Larger Version of this Image (39K GIF file)]


The binding of iron by transgenic and iron-saturated natural hLF that had been desaturated by dialysis against pH 2.5 was then studied by titration with iron. The results indicated that both desaturated transgenic and natural hLF could be completely resaturated with iron (Fig. 8). Human LF with iron at molar ratios of 0.4 and 0.8 appeared to be about 20 and 40% saturated, respectively (Fig. 8, lanes 3 and 2, cDNA transgenic hLF; lanes 7 and 6, natural hLF), whereas saturation appeared complete at an iron to hLF molar ratio of 2 to 1 (lanes 1 and 5). This indicates that both transgenic and natural hLF can bind two iron ions. Moreover, the results suggest that an hLF molecule that has bound one iron ion almost immediately binds a second one, i.e. cooperativity in iron binding of hLF N- and C-lobes appears to occur.


Fig. 8. Resaturation with iron of desaturated natural and transgenic hLF. Iron was added to desaturated cDNA transgenic hLF (lanes 1-4) and desaturated iron-saturated natural hLF (lanes 5-8) at molar ratios of 2.0, 0.8, and 0.4. Samples were diluted in nonreducing SDS sample buffer and applied to SDS-PAGE (7.5%; 10 µg of protein/lane).
[View Larger Version of this Image (68K GIF file)]


Ligand Binding Properties of Transgenic and Natural hLF

Human LF has been shown to bind to a large variety of molecules, and some of the interactions may be of physiological relevance. The results in Fig. 9 indicate that binding of transgenic hLF to DNA, LPS, and heparin is not different from that of natural hLF. The identical binding properties of the distinct hLF species to the lectins concanavalin A (30) and Ulex Europaeus agglutinin (affinity for L-fucose (31)) indicates that transgenic hLF is N-glycosylated and that the extent of fucosylation is similar as that in natural hLF. Both natural, iron-saturated hLF and transgenic hLF appeared to bind equally well to chelated iron ions. The binding was completely abrogated after selective modification of histidine residues with diethyl pyrophosphate (32), suggesting that surface-exposed histidine residues in hLF can bind metal ions (data not shown). The identical binding properties of hLF species to polyclonal rabbit anti-hLF antibody as well as to anti-hLF mAbs E1, E2, and 13.17, which bind to different conformational determinants in hLF,2 suggest that the tertiary structure of transgenic hLF is similar if not identical to that of natural hLF.


Fig. 9. Binding of purified transgenic and natural hLF to ligands immobilized onto agarose. The binding of distinct hLF species to a variety of ligands was studied by RIAs as described under "Materials and Methods." The figure shows the mean and S.D. of three separate experiments.
[View Larger Version of this Image (65K GIF file)]



DISCUSSION

Recombinant expression of hLF enables studies of the relation between protein structure and function. So far, rhLF has been expressed in non-human (22) and human cell lines (16), in yeast (20), fungi (33), tobacco plant cells (34), and in transgenic mice (18). Recombinant expression also allows large-scale production for studies of potential applications in human health care. For this reason we have genetically modified cattle to obtain expression of hLF in the bovine mammary gland (35). In this paper we demonstrate the feasibility of producing functional rhLF in the milk of transgenic mice.

The results of double antibody immunoassays (Figs. 1 and 9) and hyperimmunization of transgenic mice with natural hLF (Fig. 2) indicate that the murine and rabbit immune systems do not distinguish transgenic from natural hLF. The tolerance for natural hLF in male and female transgenic mice suggests that the expression of transgenic hLF is not completely restricted to lactation and may have occurred in other tissues during maturation. Transcriptional activity of the alpha S1 casein gene has for example been shown in the thymus (36). The strongly reduced but detectable immune response in two of the six hLF cDNA mice (2 and 8% of control) suggests that tolerance may be partially disrupted upon aggressive immunization of lines containing less powerful hLF expression vectors. This idea is supported by hyperimmunization experiments in mouse lines expressing hLF or other human proteins in milk over a wide concentration range (data not shown). The immunological identity of transgenic and natural hLF suggests that the tertiary structure and glycosylation of hLF produced in the heterologous mammary gland is very similar, if not identical, to that of natural hLF. Thus, we do not expect that administration of transgenic hLF to humans will elicit an immune response. Data on immunological identity between recombinant hLF produced in various systems (20, 22, 33, 34) to natural hLF have not been reported.

The strongly positively charged N terminus of hLF plays an important role in its bactericidal activity (37), in its binding to LPS (9) and glycosaminoglycans (38), as well as in its clearance from the circulation (39). Mono S chromatography and N-terminal sequencing (Fig. 3) revealed equal cationic properties and intact N-terminal sequences of natural and transgenic hLF.

Purified transgenic hLF appeared to be about 90% saturated with iron, whereas natural hLF was saturated for about 3% (Figs. 4 and 6). Western blots of transgenic mouse milk samples showed transgenic hLF was already saturated with iron prior to purification.2 Expression of recombinant hLF by cell lines or microorganisms also has been shown to yield hLF saturated with iron (16, 20, 22, 33) and Cu2+ (20) probably due to the presence of excess metal ions in the culture medium (16, 20). We anticipate that transgenic hLF expressed in the bovine mammary gland will show a degree of iron saturation lower than 10% based on analyses of natural hLF that had been added to cow milk.2

Transgenic hLF did not differ from natural hLF in its ability to release bound iron with decreased pH (Fig. 7), nor in its ability to bind two iron atoms concomitantly and maintain its iron-saturated conformation in the presence of SDS (Fig. 8). Identical iron binding properties imply that unsaturated transgenic and natural hLF will behave identically with respect to biological effects mediated through binding of iron, e.g. bacteriostasis (7), inhibition of hydroxyl radical formation (14), and transport of iron from the mother to the newborn (6, 27).

Analysis of untreated and enzymatically deglycosylated transgenic and natural hLF (Fig. 5) indicated that the different mobility of transgenic and natural hLF on SDS-PAGE results from differences in N-glycosylation between murine and human mammary gland epithelium. Spik et al. (4) have shown that glycosylation of lactoferrin varies between animal species, its site of production, and even at a single production site. The relative densities of the two doublets of transgenic hLF were different in cDNA and genomic transgenic hLF (Fig. 4), most likely relating to the presence of a cysteine in stead of a glycine at position 404 in cDNA transgenic hLF, since this amino acid change results in increased glycosylation at the optional glycosylation site Asn624 (16, 26).

The role of glycosylation in the biological actions of hLF is not known, but may serve to protect the molecule against rapid degradation in the gastrointestinal tract (16, 40). The identical binding properties of natural and transgenic hLF to concanavalin A and UEA I suggests that transgenic hLF like natural hLF (4) contains biantennary glycan chains. Thus, hLF expressed in the mammary gland is likely to be more similar to natural hLF than hLF produced in yeast or fungi (33).

Some of the biological activities of hLF arise from its capacity to bind to other molecules. For example, direct binding of hLF to LPS of Gram-negative bacteria with disturbance of cell wall stability has been implicated in its bactericidal action (9, 10). The interaction of hLF with the lipid A moiety of LPS (9) may account for the reduced production of cytokines upon challenge with LPS (15) as well as for the inhibition of neutrophil priming (41). The interaction with DNA of hLF translocated to the nucleus may alter the transcriptional activity of certain genes, which may contribute or explain for the immunomodulatory effects of hLF (42). The neutralization of the anticoagulant activity of heparin by hLF released from neutrophils at an infection site may serve to localize the infection through encapsulation with fibrin (43). The affinities for each of these possibly physiologically relevant ligands did not differ between transgenic hLF, iron-saturated hLF, and natural hLF (Fig. 9).

In conclusion, the similar properties of unsaturated transgenic hLF and natural hLF suggest that hLF produced in bovine milk will exert similar, if not identical, antibacterial and anti-inflammatory activities in vivo.


FOOTNOTES

*   The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§   To whom correspondence should be addressed: Pharming BV, Niels Bohrweg 11-13, 2333 CA Leiden, The Netherlands. Tel.: 31-71-5247416; Fax: 31-71-52416507; E-mail: jnuijens{at}pharming.nl.
1   The abbreviations used are: LF, lactoferrin; hLF, human lactoferrin; natural hLF, hLF from human milk; iron-saturated natural hLF, natural hLF that has completely been saturated with iron in vitro; transgenic hLF, recombinant hLF secreted in milk of transgenic mice; LPS, lipopolysaccharide; mAb, monoclonal antibody; RIA, radioimmunoassay; MES, 4-morpholineethanesulfonic acid.
2   J. H. Nuijens and P. H. C. van Berkel, unpublished observation.
3   F. R. Pieper, manuscript in preparation.

ACKNOWLEDGEMENTS

We thank Marianne Kroos (Erasmus University, Rotterdam, The Netherlands) for performing the N-terminal sequencing and Dr. Peggy Neville (Dept. of Physiology and Medicine, University of Colorado School of Medicine, Denver, CO) for critically reading the manuscript. Dr. B. Schoepen (Lichthaart, Belgium) is acknowledged for helpful discussion.


REFERENCES

  1. Metz-Boutigue, M., Jolles, J., Mazurier, J., Schoentgen, F., Legrand, D., Spik, G., Montreuil, J., and Jolles, P. (1984) Eur. J. Biochem. 145, 659-676 [Abstract]
  2. Anderson, B. F., Baker, H. M., Norris, G. E., Rice, D. W., and Baker, E. N. (1989) J. Mol. Biol. 209, 711-734 [Medline] [Order article via Infotrieve]
  3. Rey, M. W., Woloshuk, S. L., de Boer, H. A., and Pieper, F. R. (1990) Nucleic Acids Res. 18, 5288 [Medline] [Order article via Infotrieve]
  4. Spik, G., Coddeville, B., Mazurier, J., Bourne, Y., Cambillaut, C., and Montreuil, J. (1994) Adv. Exp. Med. Biol. 357, 21-32 [Medline] [Order article via Infotrieve]
  5. Johanson, B. (1960) Acta Chem. Scand. 14, 510-512
  6. Nuijens, J. H., van Berkel, P. H. C., and Schanbacher, F. L. (1996) J. Mammary Gland Biol. Neoplasia 1, 285-295 [Medline] [Order article via Infotrieve]
  7. Reiter, B., Brock, J. H., and Steel, E. D. (1975) Immunology 28, 83-95 [Medline] [Order article via Infotrieve]
  8. Arnold, R. R., Cole, M. F., and McGhee, J. R. (1977) Science 197, 263-265 [Medline] [Order article via Infotrieve]
  9. Appelmelk, B. J., An, Y. Q., Geerts, M., Thijs, B. G., de Boer, H. A., MacLaren, D. M., de Graaff, J., and Nuijens, J. H. (1994) Infect. Immun. 62, 2628-2632 [Abstract]
  10. Ellison, R. T., III, and Giehl, T. J. (1991) J. Clin. Invest. 88, 1080-1091 [Medline] [Order article via Infotrieve]
  11. Ellison, R. T., III, Luo, Q., and Reller, L. B. (1990) J. Antimicrob. Chemother. 25, 479-481 [Medline] [Order article via Infotrieve]
  12. Soukka, T., Tenovuo, J., and Lenander, L. M. (1992) FEMS Microbiol. Lett. 69, 223-228 [Medline] [Order article via Infotrieve]
  13. Petschow, B. W., and Talbott, R. D. (1991) Pediatr. Res. 29, 208-213 [Medline] [Order article via Infotrieve]
  14. Britigan, B. E., Hassett, D. J., Rosen, G. M., Hamill, D. R., and Cohen, M. S. (1989) Biochem. J. 264, 447-455 [Medline] [Order article via Infotrieve]
  15. Crouch, S. P., Slater, K. J., and Fletcher, J. (1992) Blood 80, 235-240 [Abstract]
  16. van Berkel, P. H. C., Geerts, M. E. J., van Veen, H. A., Kooiman, P. M., Pieper, F. R., de Boer, H. A., and Nuijens, J. H. (1995) Biochem. J. 312, 107-114 [Medline] [Order article via Infotrieve]
  17. Nichols, B. L., McKee, K. S., Henry, J. F., and Putman, M. (1987) Pediatr. Res. 21, 563-567 [Abstract]
  18. Platenburg, G. J., Kootwijk, E. P., Kooiman, P. M., Woloshuk, S. L., Nuijens, J. H., Krimpenfort, P. J., Pieper, F. R., de Boer, H. A., and Strijker, R. (1994) Transgenic Res. 3, 99-108 [Medline] [Order article via Infotrieve]
  19. Powell, M. J., and Ogden, J. E. (1990) Nucleic Acids Res. 18, 4013 [Medline] [Order article via Infotrieve]
  20. Liang, Q., and Richardson, T. (1993) J. Agric. Food Chem. 41, 1800-1807
  21. Panella, T. J., Liu, Y. H., Huang, A. T., and Teng, C. T. (1991) Cancer Res. 51, 3037-3043 [Abstract]
  22. Stowell, K. M., Rado, T. A., Funk, W. D., and Tweedie, J. W. (1991) Biochem. J. 276, 349-355 [Medline] [Order article via Infotrieve]
  23. Nuijens, J. H., Abbink, J. J., Wachtfogel, Y. T., Colman, R. W., Eerenberg, A. J., Dors, D., Kamp, A. J., Strack van Schijndel, R. J., Thijs, L. G., and Hack, C. E. (1992) J. Lab. Clin. Med. 119, 159-168 [Medline] [Order article via Infotrieve]
  24. Nuijens, J. H., Huijbregts, C. C. M., Van Mierlo, G. M., and Hack, C. E. (1987) Immunology 61, 387-389 [Medline] [Order article via Infotrieve]
  25. Kijlstra, A., Kuizenga, A., van der Velde, M., and van Haeringen, N. J. (1989) Curr. Eye Res. 8, 581-588 [Medline] [Order article via Infotrieve]
  26. Van Berkel, P. H. C., van Veen, H. A., Geerts, M. E. J., de Boer, H. A., and Nuijens, J. H. (1996) Biochem. J. 319, 117-122 [Medline] [Order article via Infotrieve]
  27. Cox, T. M., Mazurier, J., Spik, G., Montreuil, J., and Peters, T. J. (1979) Biochim. Biophys. Acta 588, 120-128 [Medline] [Order article via Infotrieve]
  28. Zak, B. (1958) Am. J. Clin. Pathol. 29, 590-592
  29. Mazurier, J., and Spik, G. (1980) Biochim. Biophys. Acta 629, 399-408 [Medline] [Order article via Infotrieve]
  30. Kobata, A. (1992) Eur. J. Biochem. 209, 483-501 [Abstract]
  31. Matsumoto, I., and Osawa, T. (1969) Biochim. Biophys. Acta 194, 180-189 [Medline] [Order article via Infotrieve]
  32. Mazurier, J., Leger, D., Tordra, V., Montreuil, J., and Spik, G. (1981) Eur. J. Biochem. 119, 537-543 [Medline] [Order article via Infotrieve]
  33. Ward, P. P., Piddington, C. S., Cunningham, G. A., Zhou, X., Wyatt, R. D., and Conneely, O. M. (1995) Bio/Technology 13, 498-503 [CrossRef][Medline] [Order article via Infotrieve]
  34. Mitra, A., and Zhang, Z. (1994) Plant Physiol. 106, 977-981 [Abstract/Free Full Text]
  35. Krimpenfort, P., Rademakers, A., Eyestone, W., van der Schans, A., van der Broek, S., Kooiman, P., Kootwijk, E., Platenburg, G., Pieper, F., Strijker, R., and de Boer, H. A. (1991) Bio/Technology 9, 844-847 [Medline] [Order article via Infotrieve]
  36. Grusby, M. J., Mitchell, S. C., Nabavi, N., and Glimcher, L. H. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 6897-6901 [Abstract]
  37. Bellamy, W., Takase, M., Yamauchi, K., Wakabayashi, H., Kawase, K., and Tomita, M. (1992) Biochim. Biophys. Acta 1121, 130-136 [Medline] [Order article via Infotrieve]
  38. Mann, D. M., Romm, E., and Migliorini, M. (1994) J. Biol. Chem. 269, 23661-23667 [Abstract/Free Full Text]
  39. Ziere, G. J., Bijsterbosch, M. K., and van Berkel, T. J. C. (1993) J. Biol. Chem. 268, 27069-27075 [Abstract/Free Full Text]
  40. Brines, R. D., and Brock, J. H. (1983) Biochim. Biophys. Acta 759, 229-235 [Medline] [Order article via Infotrieve]
  41. Wang, D., Pabst, K. M., Aida, Y., and Pabst, M. J. (1995) J. Leukocyte Biol. 57, 865-874 [Abstract]
  42. He, J., and Furmanski, P. (1995) Nature 373, 721-724 [CrossRef][Medline] [Order article via Infotrieve]
  43. Wu, H. F., Lundblad, R. L., and Church, F. C. (1995) Blood 85, 421-428 [Abstract/Free Full Text]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.