Main

BSDL or cholesterol esterase is an enzyme found in the pancreatic secretion of all species examined to date from fish(1, 2) to human(3). This enzyme, which has a wide range of specificity (for review, seeRef. 4), is involved in the duodenal hydrolysis of cholesteryl esters and lipid-soluble vitamin esters(5). The enzyme might act in concert with other lipolytic enzymes involved in lipid digestion such as pancreatic CDL(6) and phospholipase A2 or preduodenal lipase(7). BSDL is functionally similar and immunologically homologous to milk bile salt stimulated lipase(8, 9). Because pancreatic CDL activity is low in the infant during the first months of life(10, 11), it has been suggested that milk bile salt-stimulated lipase serves a compensatory role in fat digestion in breast-fed infants(12, 13), may promote growth in a newborn cat animal model(14), and allow better lipid absorption in low birth weight infants(15). These studies have, however, ignored pancreatic BSDL and its possible function in the duodenum of the newborn. In a recent study Ellis and Hamosh(16) described the strong similarities in both structural and functional characteristics of milk BSDL produced by the ferret and by the human, the former presenting a very high hydrolytic efficiency compared with the human enzyme(16). Therefore, use of the ferret as an animal model allows us to compare the level of BSDL in mammary gland with that of the pancreatic enzyme in the adult and in the suckling animal. In the present study, we have used the ferret to quantitate mRNA hybridizing to BSDL probes in mammary gland and adult and newborn pancreas tissues and to compare the expression of the enzyme at these sites. The overall results indicate that even though the amount of mRNA hybridizing to BSDL probes was high in newborn pancreas, the amount of enzyme expressed in this tissue was lower than that expressed in adult pancreas. The high amount of BSDL expressed in the mammary gland indicates that, in the newborn, BSDL produced in this tissue and secreted into milk might have the postulated compensatory role in fat digestion in the newborn, once activated by primary bile salts present in the duodenum.

METHODS

Biologic material. One lactating ferret (gil) and her four kits, species Mustela putoris furo (Marshall Farms, North Rose, NY) were obtained at 2 wk postpartum. The gil was over 20 wk old and weighed approximately 900 g. The four kits were all female and weighed 66-71 g.

Tissue collection. All ferrets were anesthetized with Nembutal(35 mg/kg) by intraperitoneal injection and killed by exsanguination. After a midline abdominal incision, tissues were collected into preweighed, cooled sterile plastic tubes and immediately frozen in dry ice. Ten specimens representing all lactating mammary gland sites were collected within 10 min of sacrifice; the pancreas was removed within 10-20 min in the adult and within 5 min in kits.

The stomach and its contents were removed. An incision was made along the entire length of the lesser curvature, and stomach contents were removed by gentle inversion of the stomach without trauma to the gastric mucosa. The stomach was rinsed with saline and secured to a working surface on ice. Biopsies (5 mg) were taken with pinch biopsy forceps. Collection of gastric specimens was performed according to Kirk et al.(17). Briefly, eight biopsies were taken along the greater curvature (numbered from 1 to 6), including the antrum (numbers 12 and 13), three were taken along the lesser curvature (numbers 9 to 11), and two were taken in the mid-body region (numbers 7 and 8). Biopsies were immediately placed in saline on ice and then stored at -70°C. Tissue homogenization was performed as previously described(17).

cDNA probes. The cDNA probe for BSDL was obtained by reverse transcription and polymerase chain reaction as described by Roudani et al.(18). The cDNA fragment was cloned in Puc 18 and amplified in the JM 101 Escherichia coli strain. After amplification the cDNA specific for BSDL (477 bp) was excised from the vector withPst I and EcoRI digestion. The cDNA probe for CDL(19) was a gift from Dr. B. Kerfelec and Dr. C. Chapus(CCBM-CNRS, Marseille). This cDNA clone was 1491 bp in length and was subcloned into pSG5 plasmid vector. After amplification into JM 101 E. coli strain, the complete cDNA for horse lipase was excised from the vector by EcoRI digestion. This insert was purified on agarose gel (1%), and a 591-bp probe specific for CDL was then generated by BamHI digestion. The cDNA probe specific for actin was a generous gift from Dr. R. Planels (INSERM U. 38, Marseille).

cDNA probe labeling. Inserts specific for BSDL and for CDL were purified on 1% agarose gel. Double-stranded cDNA fragments were32 P-labeled by random priming using [α-32P]dCTP (NEN, Les Ulis, France) and the Random Primed DNA-labeling (Boehringer Mannheim, Germany) kit to a specific radioactivity of 4 × 108 cpm/μg.

Northern blotting. The total RNA were extracted from frozen tissues using the method of Chirgwin et al.(20) with slight modifications, after pulverization in a chilled homogenizer in the presence of 4 M guanidinium isothiocyanate to inhibit RNases. Twenty micrograms of these total RNA suspended in 10 mM phosphate buffer (pH 7.0), glyoxal (18%), and DMSO (50%) were subjected to electrophoresis on 1% agarose gel. The RNA were then transferred to nylon membranes (Biodine A Pall-Bio-support, Portsmouth, England). The membranes were prehybridized for 3-4 h at 42°C, in 50% formamide, saline sodium citrate (2 × SSC, 1 × = NaCl, 150 mM and sodium citrate, 15 mM), SDS (1%), Denhardt's reagent (5 ×; 1 × = Ficoll, polyvinylpyrolidone, and BSA, 0.02% wt/vol each) supplemented with 200μg/mL heat-denatured herring sperm DNA. The hybridization with the labeled cDNA fragment (100 ng, 4 × 108 cpm/μg) took place overnight at 42°C in the hybridization mixture. The blots were washed three times for 5 min at room temperature with 2 × SSC, 0.1% SDS, and twice for 15 min at 60°C with 0.2 × SSC, 0.1% SDS. The filters were then autoradiographed for 24 h at -80°C using an intensifying screen.

For mRNA quantitation, total RNA were dotted onto a nitrocellulose membrane(BA 83 type, Schleicher & Schuell, Dassel, Germany) in decreasing rank amount from 10 to 0.312 μg/well. Prehybridization and hybridization were performed as above. The amount of mRNA specific for BSDL and for CDL was estimated by the dark intensity of spots on the film by a spectrophotometric method using a Dynatec MR 5000 microplate spectrophotometer. Typical hybridization lines were obtained by linear regression. From the slope of each line, it was possible to calculate the amount of specific mRNA (in arbitrary units) in the mixture of total RNA. Contents of mRNA were normalized to those in the adult pancreas.

Enzyme assays. BSDL activity was measured according to Freedet al.(21), using glycerol tri-[9,10-(N)3H]oleate as substrate buffered in gum arabic emulsion. CDL was determined at pH 8.0 according to Lowe(22) also using glycerol tri-[9,10-(N)3H]oleate. To distinguish BSDL from CDL, activities were determined with and without eserine, a powerful inhibitor of BSDL. The activity in the presence of eserine (10 mM) was subtracted from that without the inhibitor, thus giving the activity of the BSDL.

The esterolytic activity of BSDL was determined on soluble 4-nitrophenylhexanoate as previously described(2). One unit of enzyme activity corresponds to the release of 1 μmol of product(FFA or 4-nitrophenol) per min. Activities were determined always using several different amounts of the tissue homogenate and averaged.

Gastric lipase activity was determined with a glycerol tri-[9,10-(N)3H]oleate emulsion. The emulsion consisted of the radiolabeled substrate and glyceroltrioleate as carrier (200 μmol) and phosphatidylcholine (15 μmol) in 3.3 mL anhydrous glycerol. The gastric lipase assay system contained 5 mM glyceroltrioleate in 50 mM acetate buffer(pH 5.4), 7 mg of BSA, and gastric tissue homogenate in a final volume of 200μL(23). Incubations were carried out for 15 min at 37°C in a Dubnoff shaking water bath. The reaction was stopped as described by Belfrage and Vaughan(24), and the3[H]oleic acid released was separated from the triglyceride substrate by liquid-liquid partition and quantitated in aliquots (0.5 mL) of the aqueous phase, using 5 mL of Scintiverse (Fisher brand). Activity was measured in a Beckman liquid scintillation counter (LS 7500). Internal standards were used for quench corrections(16). The partition coefficient of oleic acid into the aqueous phase was determined to be 70% in this assay system and was not affected by assay conditions (pH or presence of BSA).

Protein assay. Protein concentrations were determined by the method of Lowry et al.(25) using serum albumin as standard.

BSDL protein determination. The amount of BSDL present in the samples was determined by the ELISA previously described, using polyclonal antibodies to human BSDL and human BSDL as a standard(26).

Statistical analysis. Data are expressed as mean ± SD, unless otherwise stated, for four individual kits and one gil. Mammary gland was sampled randomly at 10 sites representing the whole length, i.e. pectoral, abdominal, and inguinal regions.

RESULTS

Northern blot analyses of RNA extracted from tissues. Northern blot analyses of total RNA extracted from lactating mammary gland and adult and kit pancreas are shown in Fig. 1. In lactating mammary gland one transcript of approximately 2.3 kb was detected using the 477-bp cDNA probe specific for BSDL. A trace of a second transcript was also detected at 3.5 kb, the origin of which is unknown, but it can represent either nuclear mRNA or another species of BSDL, possibly expressed in some lactating tissue. In pancreatic tissues only one mRNA was detected at 2.2-2.3 kb, the size of which correlated with that of the transcript of the gene of BSDL. This pattern indicated that neither extracted mRNA was degraded. This was confirmed by a cDNA probe for actin, which hybridized with a unique transcript of the expected size, i.e. 2.0 kb (not shown), and suggested that the amount of mRNA specific for BSDL could be higher in the kit than in the adult pancreas and lactating mammary gland (compare lane 1, or better lane 2, with lane 3 and 4 of Fig. 1).

Figure 1
figure 1

Northern blot analyses of RNA extracted from tissues. Twenty micrograms of RNA extracted from tissues were analyzed on 1% agarose gel and transferred onto nylon membrane. (A) The 477-bp probe for BSDL was used to detect specific mRNA. RNA from kit pancreas, sample 1(lane 1) and 3 (lane 2); RNA from adult pancreas, sample 3(lane 3); RNA from lactating mammary gland, sample 6 (lane 4). (B) A 600-bp probe for actin was used to detect specific mRNA. RNA from mammary gland, sample 3 (lane 1) adult pancreas sample 7 (lane 2), and kit pancreas (pool of all specimens of RNA from kit pancreas) (lane 3).

Full size image

Quantitation of mRNA hybridizing to the BSDL probe. mRNA hybridizing to the BSDL probe (see Fig. 2, left panel, for a typical experiment) can be detected in the lactating mammary gland, in adult pancreas, and in kit pancreas. The amount of mRNA encoding for BSDL decreases in the rank order; kit pancreas > adult pancreas > lactating mammary gland (Table 1), indicating that the amount of mRNA hybridizing to the BSDL probe was about twice as high in the pancreas of the newborn than in adult tissue. As the pancreatic tissue ontologically originates from dorsal and ventral buds emerging from the duodenal endoblast, the kit pancreas was divided into four parts from the head of the pancreas(part A) to the tail (part D). Values ranging from 2.1 ± 0.3, 2.4± 0.7, 1.7 ± 0.7, and 1.9 ± 0.5, in part A, B, C, and D, respectively, indicate that there was no significant difference between these four regions of the kit pancreas.

Figure 2
figure 2

Dot-blot of RNA using probes specific for BSDL and CDL. Autoradiography of a single dot-blot of a typical hybridization experiment with the cDNA specific for BSDL (477 bp) or CDL (591 bp). Total RNA (0.15 to 10 μg) isolated from lactating mammary gland or from kit and adult pancreas was spotted onto nitrocellulose and hybridized with 32P-labeled random-primed specific cDNA probes.

Full size image
Table 1 Content of mRNA specific for BSDL and CDL of ferret lactating mammary gland, adult pancreas, and kit pancreas
Full size table

Quantitation of mRNA hybridizing to the CDL probe. mRNA hybridizing to the CDL probe was detected in the adult and kit pancreas. We were unable to detect this mRNA in lactating mammary gland (seeFig. 2, right panel, for a typical experiment). The amount of mRNA specific for CDL was greater in adult pancreas than in kit pancreas (Table 1).

Expression of BSDL in pancreatic tissues and mammary gland. The amount of BSDL present in pancreas of kits or adult and in lactating mammary gland was determined by means of an ELISA(26). Again, the amount of BSDL present in adult pancreas, which was about 4-fold higher(Table 2) than in lactating mammary gland, was compatible with the amount of mRNA specific for BSDL and with the BSDL activity present in these tissues.

Table 2 BSDL content of lactating mammary gland, adult pancreas, and kit pancreas
Full size table

These results indicate that mRNA hybridizing to BSDL probe in kit pancreas was either poorly translated into active BSDL or was highly unstable. Of significant importance is the finding that BSDL expressed in kit pancreas was also much less active, as its specific activity was about 5 to 10 times lower than in adult pancreas.

Activity of digestive lipases in the newborn ferret. The enzymatic activity of BSDL was determined in tissue homogenates of lactating mammary gland, adult pancreas, and kit pancreas, the latter numbered from 1 to 4 for the four kits studied. BSDL activity was measured either with emulsified[3H]triolein or soluble 4-nitrophenylhexanoate. As shown inTable 3, BSDL activity was markedly higher in adult pancreas than in kit pancreas, a finding incompatible with the amount of mRNA in these tissues (Table 1). The activity in lactating mammary gland and adult pancreas correlates however, quite well with the amount of mRNA in these tissues, respectively.

Table 3 Activity of BSDL and CDL in lactating mammary gland, adult pancreas, and kit pancreas
Full size table

CDL activity could not be detected in kit pancreas and adult lactating mammary gland but was present at high levels in adult pancreas. In this case the level of mRNA specific for CDL in each of these tissues paralleled the enzyme activity in the corresponding tissue.

Because the stomach plays an important role in fat digestion shortly after birth(13, 17, 23), we have determined the lipase activity in gastric mucosa of newborn and adult ferret(Table 4). Biopsy specimens were taken at various sites of the gastric mucosa(17). In the adult, gastric lipase activity was higher in the gastric body than in the antrum. Lipase activity was lower in the gastric body of kits, compared with adults. Lipase activity was much lower in the antrum than in the body, and was not different as a function of age.

Table 4 Lipase activity in the gastric mucosa of the suckling ferret
Full size table

DISCUSSION

In the present study, transcriptions of the BSDL gene and of the CDL gene were detected by means of specific probes. The BSDL gene was actively transcribed in lactating mammary gland, in adult pancreas and in kit pancreas. A transcription product of the CDL gene was detected in the adult pancreas, but could not be detected in the lactating mammary gland.

The amount of mRNA specific for CDL was higher in adult than in kit pancreas. On the contrary, the amount of mRNA specific for BSDL was significantly higher in the newborn pancreas compared with the adult pancreas. To correlate the transcription level of the BSDL gene with the translation of mRNA issued from this gene, we also quantitated BSDL by an ELISA technique specific for BSDL. Data show that the translation of mRNA specific for BSDL was significantly lower in kit pancreas than in adult pancreas as indicated by the amount of BSDL present in kit pancreas compared with adult pancreas. This suggests that the mRNA specific for BSDL was either less stable in kit than in adult pancreas, or more likely, that the translation of this mRNA was not as effective.

A second significant observation concerns the activity of the BSDL expressed by kit pancreas. Even though the amount of BSDL expressed in kit pancreas (0.2 ± 0.1 mg of BSDL/mg of protein) was comparable to the amount present in lactating mammary gland (≈0.3 mg of BSDL/mg of protein), the enzyme-specific activity was about three times lower. When compared with the adult pancreas, the amount of BSDL present in kit pancreas was approximately five times lower, whereas its specific activity was 5 to 10 times lower. These data suggest that the BSDL expressed in kit pancreas (at least in three out of four kits) was less active than the enzyme expressed in adult ferret pancreas. We have recently described a fetal glycoform of BSDL(referred to as FAP(27), which was expressed in fetal human pancreas as well as in human pancreatic tumors(28, 29). FAP displayed the same specificity as BSDL; however, its catalytic efficiency was only 10% of that of BSDL(27). In the human, FAP was mainly expressed in the fetus and only poorly in the adult(30), whereas the amount of BSDL increased with time from parturition to adult age. One difference between FAP and BSDL is the presence of a carbohydrate-dependent epitope on FAP that is recognized by the MAb J28(31). Nevertheless, differences in amino acid composition between BSDL and FAP were also noted(27). This difference lies in amino acids represented mostly in the C-terminal part of the proteins(27). As the cDNA probe used in this study hybridizes with the sequence encoding the central part of BSDL, one may assume that it hybridizes also with the mRNA, the translation of which would give FAP. These points indicate that the amount of mRNA determined in kit pancreas may reflect a cross-hybridization with FAP mRNA, explaining the higher amount of mRNA detected in kit pancreas than adult pancreas. As mentioned above, FAP is much less active than BSDL, also antibodies specific of BSDL used in the ELISA are probably less reactive with FAP (our unpublished observation), this gives a logical (but still hypothetical) explanation to the low esterolytic activity and low amount of BSDL protein determined in kit pancreas compared with adult pancreas. With regard to CDL, it is possible that the probe used here may also cross-react with mRNA of the hPLRP-1 and hPLRP-2 described by Giller et al.(32). The expression of these two hPLRP-1 and hPLRP-2 lipases related mRNA is very low compared with that of CDL (4 times and 24 times lower, respectively). In the rat, the pattern of CDL and lipase-related proteins 1 and 2 mRNA expression suggests that regulation of PLRP is coordinate as opposed to the discoordinate regulation of the PLRP and CDL mRNA(33). The level of all these mRNA is low in fetal rat pancreas and PLRP mRNA levels rise just before birth and fall to 5-10% to adult levels by 14-21 d. During this time the CDL mRNA level increases to the adult level(33). As the same situation possibly exists in ferrets, our CDL probe may also hybridize with ferret PLRP mRNA. Taken together these data explain why CDL activity cannot be determined in 2-wk kit pancreas, whereas mRNA hybridizing to the CDL probe were detected in significant amount.

As in dog(17) and human(34), the preduodenal lipase activity of the ferret is localized mainly in the body of the stomach; the antrum has only negligible lipase levels. In other species, preduodenal lipase activity is localized in the cardia area as in rabbit(35) or is diffusely distributed throughout the entire mucosa as in guinea pig and baboon(35). The data presented show that preduodenal lipase activity in the gastric mucosa of the newborn ferret is similar in location to that of adults. The level of gastric lipase activity was about two times lower in the newborn than adult ferret. This is in marked contrast to the developmental pattern of pancreatic lipases, the synthesis and secretion of which are significantly lower in the newborn as reported here for the ferret and previously for the human and the rat(10, 30, 33, 36, 37). The level of gastric lipase is compatible with an effective hydrolysis of milk fat globules in the newborn stomach.

The data reported in this study, namely markedly reduced message and expression of CDL and low expression of BSDL, in spite of adequate message in the newborn pancreas, strengthen the assumption that milk BSDL is an important compensatory enzyme for fat digestion in suckling primates and carnivores. The relatively low contribution of pancreatic lipases to the digestive process, during a period of high fat consumption, indicates that the sucklings of these species depend upon gastric lipase and milk BSDL for adequate digestion of milk fat. Initiation of hydrolysis of milk fat by gastric lipase is a prerequisite for the subsequent action of BSDL, which cannot access the triglycerides within the core of milk fat globules before partial lipolysis in the stomach(38–40). The relative contribution of exogenous (milk) and endogenous (gastric and pancreatic) lipases to fat digestion in the newborn varies among species(40).