428
Gahretal.
washed twice for nonspecific sion contained
in HBSS, stained with Wright stain tested esterase activity [9, 10]. The cell suspenless than 1% contaminating PMN.
Cultures of Blood Macrophages
Monocytes
Monocyte derived macrophages serum-free teflon bag system for elsewhere [11].
Preparation
to Generate were cultivated in a 7 days as described
of Stimuli
Chemotaxis. Casein (Merck) was used in a final concentration of 5 mg/ml in HBSS containing 0.1% human albumin, the pH adjusted to pH 7.3. FMLP (Sigma) was dissolved in absolute ethanol to lO M, then diluted with HBSS to a final concentration of i07 M, containing 0. 1% human albumin. Zymosan activated serum was prepared by incubation of 1 mg zymosan A (Sigma) with 1 ml AB serum at 37#{176}C for 45 mm. After 30 mm at 56#{176}C the zymosan particles were centrifuged for 15 mm at 1,800g. To the supernatant 4 ml HBSS was added with 0.1% human albumin and aliquots stored at -30#{176}Cuntil use. Superoxide production and degranulation. Phorbol myristate acetate (Sigma) dissolved in dimethyl sulfoxide was used in a final concentration of 1 ig/ml, 100 ng/ml and 10 ng/ml. FMLP in a concentration of l0 M in dimethyl sulfoxide was diluted to a final concentration of l0 M in the test. Zymosan A was opsonized by incubation of 30 mg with 2 ml AB serum at 37#{176}C for 15 mm. This suspension was centrifuged 10 mm at 800g at 25#{176}C.The pellet was washed with HBSS and resuspended in 2 ml HBSS. The final concentration was 1.5 mg/ml for the determination of superoxide and 6 mg/ml in the degranulation assays.
PMN Aggregation The aggregation of PMN was measured as described previously [12] in an aggregometer; an increase of the light transmission at 600 nm indicated an increase of aggregation as measured on an XY recorder.
Cell Motility Cell motility was determined in a 48-well microchemotaxis chamber (Neuroprob, Cabin John, MD, USA). The upper cell compartment and the lower compartment were divided by a nitrocellulose filter, 150 im (Sartorius, Gottingen, FRG) with a pore size of 3 p.m for PMN and 5 p.m for monoytes. Absolute cell count in the upper compartment was 6 x i04 cells in a volume of 60 p.1 HBSS with 0.1% human albumin (Behring, Marburg, FRG). For monocytes the same concentration
was used. Human albumin was always present in the lower compartment (0. 1 % final concentration) irrespective of the chemotactic factor used. After 1 hr at 37#{176}C in 5% CO2 the filters were stained as described by Wilkinson [ 13] The depth into the filter to which the leading front of cells had migrated was measured by microscopic examination. .
Degranulation PMN (1 x l07/ml) in HBSS were preincubated in a total volume of 1 .5 ml without or with lactoferrin (50-500 p.g/ml) for 30 mm at 37#{176}C. Then 200 p.1 was removed, centrifuged 2 mm at 12,000g, and the supernatant stored at -40#{176}C.The pellet was sonicated (three times 30 sec, 50 W on ice) to determine the total content of granule proteins. Then degranulation was induced by the addition of phorbol myristate acetate ( 1 p.g/ml) or opsonized zymosan (6 mg/ml). After 10, 20, and 30 mm aliquots were removed, centrifuged as mentioned above, and the supernatants stored at -40#{176}C. In the supernatants and the sonicated pellets lysozyme was determined by measurement of lysis of Micrococcus lvsodeikticus using the test kit Testomar-lysozyme of Behring (Marburg, FRG) [141. 3-Glucuronidase was quantitated by liberation of nitrophenol from nitrophenylglucuronide [14]. All assays were performed in duplicate. The activity of lysozyme and 3-g1ucuronidase was expressed as absorbance units (A/106 cells/mm and/2 hr, respectively). The total cellular content of both enzymes in absorbance units has been published elsewhere [14]. The lactate dehydrogenase activity in the supernatants did not exceed 2.5% of the total cellular content during the incubation.
Release
of Superoxide
Release of superoxide was measured by continuous (PMN) or discontinuous (monocytes and macrophages) determination of superoxide dismutase-inhibitable reduction of cytochrome c at 37#{176}C. The assay mixture contained in a final volume of 1 ml: 50 nmol cytochrome c (Sigma) and 1 X 106 PMN or 1 X l0 monocytes or macrophages in HBSS without or with 33 U superoxide dismutase (from bovine erythrocytes. Sigma). The reaction was started by the addition of phorbol myristate acetate (100 p.1), opsonized zymosan (100 p.1), or FMLP (10 p.1) in the concentrations mentioned above. For PMN the reaction rate was determined in the linear part of the absorbance change at 550 nm (usually from 2 to 6 mm). The superoxide production of monocytes and macrophages was expressed as nmol/105 cells/hr and that of PMN as nmo1/l0 cells/mm. The amount of superoxide was calculated on the basis, that a change in absorbance of 1.0 corresponds to 47.4 nmol superoxide. Each reaction was run in duplicate or triplicate.
Lactoferrin and PMN Function Oxygen
Consumption
Bacterial
FMLP
PMA
0.9
.
E LI, Lfl
0.7 1.3
>%
U)
Other Assays
C
All buffers and reagents used were determined to be free from detectable bacterial endotoxin by the limulus amebocyte lysate assay (Sigma). With this assay it is possible to detect endotoxin concentrations as low as 0. 1 ng/mI. Statistical analysis was performed by Student’s t test for unpaired samples.
a
1.1
0.9 U
ci
Production
0.7
0
1.3
Consumption
Preincubation of PMN with lactoferrin for 30 mm at 37#{176}C resulted in an increased production rate of superoxide (Fig. 1A). This was observed for all stimuli used (phorbol myristate acetate. FMLP, opsonized zymosan), regardless of whether PMN were preincubated in ironpoor or in iron-saturated lactoferrin. The greatest increase of superoxide production was achieved after FMLP stimulation with lactoferrin saturated with iron (189 ± 11.9% of the superoxide production of PMN incubated without lactoferrin) (Table 1). Lowering the concentration of phorbol myristate acetate to 100 and 10 ng/ml also resulted in an increased superoxide production (Table 1). The priming effect of
I,,-.
1.1
0.9
and Oxygen
FMLP PMA
of Lactoferrin
Human lactoferrin was isolated from milk of healthy mothers as described by Sawatzki and Kubanek [18] and Leclercq et al. [19] with the following steps. The whey was iron saturated by using Fe3 -nitrilotriacetate reagent [20] after centrifugation, applied to CM-Sepharose CL6B, and eluted with a salt gradient. Further purification was achieved by DEAE Sepharose CL-6B chromatography with a salt gradient and then purified on a heparin Sepharose column again eluted by a salt gradient with NaCI. The concentrated eluates were freeze-dried for storage. Iron poor lactoferrin was prepared by dialysis against EDTA phosphate buffer pH 4. The purity of the lactoferrin was checked by SDS polyacrylamide electrophoresis which showed only one band.
RESULTS Superoxide
PMA
1.1
Killing
The assay of bacterial killing was based on the pour-plate method of Quie et al. [ 16] with a modification published recently [ I 7] The number of viable bacteria was determined at 0, 30, 60, and 90 mm incubation of phagocytes and Staphylococcus aureus.
Preparation
FMLP
1.3
Oxygen consumption was determined polarigraphically using a Clark oxygen electrode as described by Weening et al. [15].
429
0.7
‘“E’
0
2
mm
4
after
addition
0
2
4
6
8
of stimulus
Fig. 1.
Superoxide production of PMN preincubated without or with (----------) 0.5 mg/mI iron-saturated lactoferrin for 30 mm in different buffers: (A) HBSS, (B) HBSS with 0.1% human albumin, and (C) HBSS with plasma (50% v/v). Final concentration of FMLP was 10 M and of phorbol myristate acetate 0.1 ig/ml. The results of one of three representative experiments are shown.
(-)
lactoferrin on superoxide production could also be observed if PMN were in HBSS containing human albumin and also in plasma (Fig. lB and C). The increase of superoxide production after addition of lactoferrin was dose dependent; it was highest at lactoferrin concentrations of 500 p.g/ml (= 6.25 p.mol/liter) and not detectable below 100 p.g/ml (Fig. 2). We also examined whether the iron saturation of lactofemn had some influence on the superoxide production enhancing effect. A slight difference was observed only after FMLP stimulation (P < 0. 10). If lactoferrin was added immediately before stimulation we observed the same results. Under conditions which resulted in an increased superoxide production of lactoferrin-treated PMN, we could also demonstrate an increase of oxygen consumption for all stimuli used. Incubation of PMN with lactoferrin (for 30 mm) followed by washing the cells with the purpose to remove all the lactoferrin did not abolish the superoxide produc-
430
Gahretal. TABLE 1. Superoxlde Production of PMN, Influence of Lactoferrln (0.5 mg/mI)a
Monocytes,
and Macrophages
Under
the
Iron
saturated
Lactoferrin Iron Polymorphonuclear leukocytes Phorbol myristate acetate 1 g/m1 (n = 19) lOO ng/ml (n = 6) 10 ng/ml (n = 6) FMLP(n 17) Opsonized zymosan (n = 18) Monocytes Phorbol myristate acetate (n = Macrophages Phorbol myristate acetate (n = alhe
superoxide
production
poor
150 ± 10 131 ± 11.4 137 ± 12.6 155 ±4.3 134 ± 5.5
131 125 130 189± 139
± 5.1* ± 12.7* ± 13.6* 11.9t ± 11.4*
12)
103
± 14.1
97 ± 15.0*
7)
109
± 13.1
108 ± 177*
of PMN,
monocytes,
and
macrophages
is given
as the percentage
of
those of cells not preincubated with lactoferrin. The absolute amounts of superoxide produced by cells not preincubated with lactoferrin were: PMN (nmol/min/106 cells) 8.7 ± 1.7; 7.7 ± 1.6; 5.6 ± 1.9 (phorbol myristate acetate 1 g/ml; 100 ng/ml; 10 ng/ml), 6.4 ± 2.0 (FMLP), 4.6 ± 1.8 (opsonized zymosan); monocytes and macrophages (nmol/hr/105 cells) 32.4 ± 2.5 and 12.4 ± 3.9 (phorbol myristate acetate, I j.tg/ml). The mean ± SEM are given. For PMN all values obtained with lactoferrin-preincubated cells were statistically significantly higher (at least P < 0.005) than those obtained with control cells. Statistics (between iron-poor and iron-saturated lactoferrin): *not significant; tP < 0.10.
0)
200
C U
served, e.g. the cells released the same amount of superoxide as PMN preincubated only in HBSS (Table 2). Addition of human transferrin (from 10 p.g/ml to 1 mg/ml final concentration) and of FeCl3 (up to 0.2 mmol/liter) and Fe citrate (up to 0.2 mmol/liter) had no superoxide production enhancing effect (Table 2). PMN incubated with lactoferrin alone (i.e., without stimulation), either iron saturated or iron poor, did not result in any measurable superoxide production. Lactoferrin itself did not interfere with the superoxide measuring system as measured by the xanthine-xanthine oxidase system. In contrast to PMN monocytes and monocyte-derived ,
a
175
0.
150
125
100 LL I
0
50
100
200
300
400
500
Lactoferrin
I
600
I
700
pg/mi
Fig. 2. Increase of FMLP-induced superoxide production of PMN preincubated with different concentrations of iron-saturated LF. The results are expressed as the percentage of superoxide production of PMN incubated without LF. Data from one representative experimental are given; only PMN from one person have been used in this experiment.
tion-enhancing effect compared to control cells (incubated without lactoferrin) treated in the same manner (Table 2). To test further the specificity of the superoxide production-enhancing effect of lactoferrin we added to the preincubation mixture together with lactoferrin a polyclonal antibody to human lactoferrin [rabbit anti-human lactoferrin (Sigma) diluted to a final concentration of 1:80, 1:400, 1:800]. After this, the superoxide production-enhancing effect of lactoferrin could not be ob-
macrophages (cultured for 7 days) were not influenced by lactoferrin to produce more superoxide after stimulation than control cells (data shown only for phorbolmyristate acetate) regardless of the stimulus used (Table 1).
Cell Motility The most pronounced effect of lactoferrin on PMN function was that on cell motility. In the presence of lactoferrin PMN exhibited an increase of their random motility as measured in a Boyden filter system. As shown in Figure 3 PMN had a motility 2.5- to 3-fold that of control cells. With iron-poor lactoferrin the cells moved faster than with iron-saturated lactoferrin (P
Gahretal.
washed twice for nonspecific sion contained
in HBSS, stained with Wright stain tested esterase activity [9, 10]. The cell suspenless than 1% contaminating PMN.
Cultures of Blood Macrophages
Monocytes
Monocyte derived macrophages serum-free teflon bag system for elsewhere [11].
Preparation
to Generate were cultivated in a 7 days as described
of Stimuli
Chemotaxis. Casein (Merck) was used in a final concentration of 5 mg/ml in HBSS containing 0.1% human albumin, the pH adjusted to pH 7.3. FMLP (Sigma) was dissolved in absolute ethanol to lO M, then diluted with HBSS to a final concentration of i07 M, containing 0. 1% human albumin. Zymosan activated serum was prepared by incubation of 1 mg zymosan A (Sigma) with 1 ml AB serum at 37#{176}C for 45 mm. After 30 mm at 56#{176}C the zymosan particles were centrifuged for 15 mm at 1,800g. To the supernatant 4 ml HBSS was added with 0.1% human albumin and aliquots stored at -30#{176}Cuntil use. Superoxide production and degranulation. Phorbol myristate acetate (Sigma) dissolved in dimethyl sulfoxide was used in a final concentration of 1 ig/ml, 100 ng/ml and 10 ng/ml. FMLP in a concentration of l0 M in dimethyl sulfoxide was diluted to a final concentration of l0 M in the test. Zymosan A was opsonized by incubation of 30 mg with 2 ml AB serum at 37#{176}C for 15 mm. This suspension was centrifuged 10 mm at 800g at 25#{176}C.The pellet was washed with HBSS and resuspended in 2 ml HBSS. The final concentration was 1.5 mg/ml for the determination of superoxide and 6 mg/ml in the degranulation assays.
PMN Aggregation The aggregation of PMN was measured as described previously [12] in an aggregometer; an increase of the light transmission at 600 nm indicated an increase of aggregation as measured on an XY recorder.
Cell Motility Cell motility was determined in a 48-well microchemotaxis chamber (Neuroprob, Cabin John, MD, USA). The upper cell compartment and the lower compartment were divided by a nitrocellulose filter, 150 im (Sartorius, Gottingen, FRG) with a pore size of 3 p.m for PMN and 5 p.m for monoytes. Absolute cell count in the upper compartment was 6 x i04 cells in a volume of 60 p.1 HBSS with 0.1% human albumin (Behring, Marburg, FRG). For monocytes the same concentration
was used. Human albumin was always present in the lower compartment (0. 1 % final concentration) irrespective of the chemotactic factor used. After 1 hr at 37#{176}C in 5% CO2 the filters were stained as described by Wilkinson [ 13] The depth into the filter to which the leading front of cells had migrated was measured by microscopic examination. .
Degranulation PMN (1 x l07/ml) in HBSS were preincubated in a total volume of 1 .5 ml without or with lactoferrin (50-500 p.g/ml) for 30 mm at 37#{176}C. Then 200 p.1 was removed, centrifuged 2 mm at 12,000g, and the supernatant stored at -40#{176}C.The pellet was sonicated (three times 30 sec, 50 W on ice) to determine the total content of granule proteins. Then degranulation was induced by the addition of phorbol myristate acetate ( 1 p.g/ml) or opsonized zymosan (6 mg/ml). After 10, 20, and 30 mm aliquots were removed, centrifuged as mentioned above, and the supernatants stored at -40#{176}C. In the supernatants and the sonicated pellets lysozyme was determined by measurement of lysis of Micrococcus lvsodeikticus using the test kit Testomar-lysozyme of Behring (Marburg, FRG) [141. 3-Glucuronidase was quantitated by liberation of nitrophenol from nitrophenylglucuronide [14]. All assays were performed in duplicate. The activity of lysozyme and 3-g1ucuronidase was expressed as absorbance units (A/106 cells/mm and/2 hr, respectively). The total cellular content of both enzymes in absorbance units has been published elsewhere [14]. The lactate dehydrogenase activity in the supernatants did not exceed 2.5% of the total cellular content during the incubation.
Release
of Superoxide
Release of superoxide was measured by continuous (PMN) or discontinuous (monocytes and macrophages) determination of superoxide dismutase-inhibitable reduction of cytochrome c at 37#{176}C. The assay mixture contained in a final volume of 1 ml: 50 nmol cytochrome c (Sigma) and 1 X 106 PMN or 1 X l0 monocytes or macrophages in HBSS without or with 33 U superoxide dismutase (from bovine erythrocytes. Sigma). The reaction was started by the addition of phorbol myristate acetate (100 p.1), opsonized zymosan (100 p.1), or FMLP (10 p.1) in the concentrations mentioned above. For PMN the reaction rate was determined in the linear part of the absorbance change at 550 nm (usually from 2 to 6 mm). The superoxide production of monocytes and macrophages was expressed as nmol/105 cells/hr and that of PMN as nmo1/l0 cells/mm. The amount of superoxide was calculated on the basis, that a change in absorbance of 1.0 corresponds to 47.4 nmol superoxide. Each reaction was run in duplicate or triplicate.
Lactoferrin and PMN Function Oxygen
Consumption
Bacterial
FMLP
PMA
0.9
.
E LI, Lfl
0.7 1.3
>%
U)
Other Assays
C
All buffers and reagents used were determined to be free from detectable bacterial endotoxin by the limulus amebocyte lysate assay (Sigma). With this assay it is possible to detect endotoxin concentrations as low as 0. 1 ng/mI. Statistical analysis was performed by Student’s t test for unpaired samples.
a
1.1
0.9 U
ci
Production
0.7
0
1.3
Consumption
Preincubation of PMN with lactoferrin for 30 mm at 37#{176}C resulted in an increased production rate of superoxide (Fig. 1A). This was observed for all stimuli used (phorbol myristate acetate. FMLP, opsonized zymosan), regardless of whether PMN were preincubated in ironpoor or in iron-saturated lactoferrin. The greatest increase of superoxide production was achieved after FMLP stimulation with lactoferrin saturated with iron (189 ± 11.9% of the superoxide production of PMN incubated without lactoferrin) (Table 1). Lowering the concentration of phorbol myristate acetate to 100 and 10 ng/ml also resulted in an increased superoxide production (Table 1). The priming effect of
I,,-.
1.1
0.9
and Oxygen
FMLP PMA
of Lactoferrin
Human lactoferrin was isolated from milk of healthy mothers as described by Sawatzki and Kubanek [18] and Leclercq et al. [19] with the following steps. The whey was iron saturated by using Fe3 -nitrilotriacetate reagent [20] after centrifugation, applied to CM-Sepharose CL6B, and eluted with a salt gradient. Further purification was achieved by DEAE Sepharose CL-6B chromatography with a salt gradient and then purified on a heparin Sepharose column again eluted by a salt gradient with NaCI. The concentrated eluates were freeze-dried for storage. Iron poor lactoferrin was prepared by dialysis against EDTA phosphate buffer pH 4. The purity of the lactoferrin was checked by SDS polyacrylamide electrophoresis which showed only one band.
RESULTS Superoxide
PMA
1.1
Killing
The assay of bacterial killing was based on the pour-plate method of Quie et al. [ 16] with a modification published recently [ I 7] The number of viable bacteria was determined at 0, 30, 60, and 90 mm incubation of phagocytes and Staphylococcus aureus.
Preparation
FMLP
1.3
Oxygen consumption was determined polarigraphically using a Clark oxygen electrode as described by Weening et al. [15].
429
0.7
‘“E’
0
2
mm
4
after
addition
0
2
4
6
8
of stimulus
Fig. 1.
Superoxide production of PMN preincubated without or with (----------) 0.5 mg/mI iron-saturated lactoferrin for 30 mm in different buffers: (A) HBSS, (B) HBSS with 0.1% human albumin, and (C) HBSS with plasma (50% v/v). Final concentration of FMLP was 10 M and of phorbol myristate acetate 0.1 ig/ml. The results of one of three representative experiments are shown.
(-)
lactoferrin on superoxide production could also be observed if PMN were in HBSS containing human albumin and also in plasma (Fig. lB and C). The increase of superoxide production after addition of lactoferrin was dose dependent; it was highest at lactoferrin concentrations of 500 p.g/ml (= 6.25 p.mol/liter) and not detectable below 100 p.g/ml (Fig. 2). We also examined whether the iron saturation of lactofemn had some influence on the superoxide production enhancing effect. A slight difference was observed only after FMLP stimulation (P < 0. 10). If lactoferrin was added immediately before stimulation we observed the same results. Under conditions which resulted in an increased superoxide production of lactoferrin-treated PMN, we could also demonstrate an increase of oxygen consumption for all stimuli used. Incubation of PMN with lactoferrin (for 30 mm) followed by washing the cells with the purpose to remove all the lactoferrin did not abolish the superoxide produc-
430
Gahretal. TABLE 1. Superoxlde Production of PMN, Influence of Lactoferrln (0.5 mg/mI)a
Monocytes,
and Macrophages
Under
the
Iron
saturated
Lactoferrin Iron Polymorphonuclear leukocytes Phorbol myristate acetate 1 g/m1 (n = 19) lOO ng/ml (n = 6) 10 ng/ml (n = 6) FMLP(n 17) Opsonized zymosan (n = 18) Monocytes Phorbol myristate acetate (n = Macrophages Phorbol myristate acetate (n = alhe
superoxide
production
poor
150 ± 10 131 ± 11.4 137 ± 12.6 155 ±4.3 134 ± 5.5
131 125 130 189± 139
± 5.1* ± 12.7* ± 13.6* 11.9t ± 11.4*
12)
103
± 14.1
97 ± 15.0*
7)
109
± 13.1
108 ± 177*
of PMN,
monocytes,
and
macrophages
is given
as the percentage
of
those of cells not preincubated with lactoferrin. The absolute amounts of superoxide produced by cells not preincubated with lactoferrin were: PMN (nmol/min/106 cells) 8.7 ± 1.7; 7.7 ± 1.6; 5.6 ± 1.9 (phorbol myristate acetate 1 g/ml; 100 ng/ml; 10 ng/ml), 6.4 ± 2.0 (FMLP), 4.6 ± 1.8 (opsonized zymosan); monocytes and macrophages (nmol/hr/105 cells) 32.4 ± 2.5 and 12.4 ± 3.9 (phorbol myristate acetate, I j.tg/ml). The mean ± SEM are given. For PMN all values obtained with lactoferrin-preincubated cells were statistically significantly higher (at least P < 0.005) than those obtained with control cells. Statistics (between iron-poor and iron-saturated lactoferrin): *not significant; tP < 0.10.
0)
200
C U
served, e.g. the cells released the same amount of superoxide as PMN preincubated only in HBSS (Table 2). Addition of human transferrin (from 10 p.g/ml to 1 mg/ml final concentration) and of FeCl3 (up to 0.2 mmol/liter) and Fe citrate (up to 0.2 mmol/liter) had no superoxide production enhancing effect (Table 2). PMN incubated with lactoferrin alone (i.e., without stimulation), either iron saturated or iron poor, did not result in any measurable superoxide production. Lactoferrin itself did not interfere with the superoxide measuring system as measured by the xanthine-xanthine oxidase system. In contrast to PMN monocytes and monocyte-derived ,
a
175
0.
150
125
100 LL I
0
50
100
200
300
400
500
Lactoferrin
I
600
I
700
pg/mi
Fig. 2. Increase of FMLP-induced superoxide production of PMN preincubated with different concentrations of iron-saturated LF. The results are expressed as the percentage of superoxide production of PMN incubated without LF. Data from one representative experimental are given; only PMN from one person have been used in this experiment.
tion-enhancing effect compared to control cells (incubated without lactoferrin) treated in the same manner (Table 2). To test further the specificity of the superoxide production-enhancing effect of lactoferrin we added to the preincubation mixture together with lactoferrin a polyclonal antibody to human lactoferrin [rabbit anti-human lactoferrin (Sigma) diluted to a final concentration of 1:80, 1:400, 1:800]. After this, the superoxide production-enhancing effect of lactoferrin could not be ob-
macrophages (cultured for 7 days) were not influenced by lactoferrin to produce more superoxide after stimulation than control cells (data shown only for phorbolmyristate acetate) regardless of the stimulus used (Table 1).
Cell Motility The most pronounced effect of lactoferrin on PMN function was that on cell motility. In the presence of lactoferrin PMN exhibited an increase of their random motility as measured in a Boyden filter system. As shown in Figure 3 PMN had a motility 2.5- to 3-fold that of control cells. With iron-poor lactoferrin the cells moved faster than with iron-saturated lactoferrin (P
