Wasting, pre-pubescent protein-energy malnutrition (PEM)⁵ consistently depresses the ability to generate antibody-mediated protection for the wet mucosae (Chandra 1991). In contrast, the influence of PEM on antibody responses within the deep tissues, i.e., systemic humoral immunity, seems much less predictable (Chandra 1991). No clear explanation is available in relation to the apparent difference in sensitivity to wasting PEM on the part of mucosal and systemic immunoglobulin (Ig)-producing systems. Recent evidence reveals that, up to the terminal differentiation of Ig-containing cells, the systemic and mucosal Ig-producing systems exhibit similarly remarkable resistance to wasting PEM (Ha et al. 1996). These results focus attention on the translocation of mucosal Ig into mucous secretions as a process, unique to the mucosal antibody response, which may prove particularly sensitive to PEM (Ha et al. 1996). Others have suggested previously, on the basis of blood plasma and mucous Ig concentrations (McMurray et al. 1977), that PEM may reduce translocation of mucosal Ig onto epithelial surfaces.

Mucosal Ig-secreting cells are located within the subepithelial loose connective tissue, and the majority of these cells produce antibody of IgA class, e.g., at least 80% in the intestinal lamina propria (Ha et al. 1996). Transport of IgA onto mucosal surfaces is mediated by the polymeric immunoglobulin receptor (pIgR), a transmembrane protein found on the basolateral membrane of intestinal and other mucosal epithelial cells in mammalian species as well as on the sinusoidal membrane of the hepatocyte of rodents and, perhaps, of humans (Kerr 1990). The epithelial pIgR is considered important mainly in the transport of locally produced IgA, whereas the hepatic pIgR is considered to function primarily in the transport of IgA from the blood (Kerr 1990). Important species differences exist in the relative contribution of epithelial and hepatic transport to the total quantity of IgA found in intestinal mucus. In humans, epithelial transport is thought to be quantitatively the more important (Daniels and Schmucker 1987, Kerr 1990), whereas rats represent the opposite extreme (Lemaitre-Coelho et al. 1978), and mice are an intermediate case more similar to humans (Delacroix et al. 1985).

Uniquely among plasma membrane receptors, the pIgR is neither recycled nor degraded intracellularly after binding to its ligand (IgA), but a portion of the receptor is released in a complex with the ligand (Kerr 1990). The ligand-bound fragment is derived from the extracellular domain of the pIgR and is designated secretory component (SC), and the molecular complex released onto the mucosal surface is referred to as secretory IgA (Kerr 1990). In addition, both hepatocytes and extrahepatic epithelial cells seem to release unbound SC constitutively (Kerr 1990).

Watson et al. (1985) found low concentrations of unbound SC in lacrimal secretions of wasting children, and Sullivan et al. (1993) reported corroborating results in a study of lacrimal, salivary and intestinal fluids of malnourished weanling rats. Following directly from these results, the primary objective of the present investigation was to determine the concentration and total quantity of the intestinal and hepatic pIgR in mice subjected to a protocol of experimental PEM that produces low concentrations of secretory IgA in intestinal mucous secretions. The broad hypothesis emanating from this investigation, when considered together with a related study of the mucosal IgA-producing effector compartment in wasting disease (Ha et al. 1996), is that expression of the pIgR is a particularly sensitive aspect of mucosal humoral immunity in PEM. The focus of this study was on PEM in its most debilitating forms because of the need to improve understanding of the basic immunobiology of this particularly challenging clinical condition. The low protein diet used in this investigation has an extreme protein-to-energy ratio relative to diets associated with wasting PEM in humans, but the experimental protocol was designed with the primary purpose of duplicating critical features of the human disease (Ha et al. 1996).

MATERIALS AND METHODS

The liver was weighed, cut into small pieces and prepared according to the method of Perez et al. (1989) with modification. Briefly, the organ was homogenized using a Caframo homogenizer (Wiarton, ON) at a setting of 7 (Stirrer type R2R1-64). Either 10 volumes (Groups B and C) or 20 volumes (Group LP) of cold (4°C) buffer, pH 7.4, was used containing 10 mmol/L TES [N-tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid, Sigma Chemical], 0.2 mol/L sucrose, 1 mmol/L MgCl₂, 12.5 mmol/L benzamidine and 21 mmol/L leupeptin. A small sample of liver homogenate was reserved for assay of protein concentration. The remainder of the homogenate was centrifuged at 1000 × g for 10 min at 4°C, after which the supernatant was ultracentrifuged at 100,000 × g for 60 min at 4°C. The resulting pellet was resuspended in TES-sucrose buffer containing 10 mL/L Triton X-100, 12.5 mmol/L benzamidine, 21 mmol/L leupeptin and 1 mmol/L phenylmethylsulfonyl fluoride (Boehringer Mannheim Canada, Laval). This procedure yielded a crude particulate fraction, rather than a pure membrane fraction, but was adopted in order to maximize recovery of pIgR. The crude particulate fraction was stored at 80°C.

Intestinal tissue, washed free of contents using cold (4°C) PBS, was blotted on filter paper and weighed. A crude particulate fraction of intestine was prepared and stored as described for the liver samples, except that the intestine was initially homogenized for 20 s by means of a homogenizing apparatus (Ultra-Turrax®, Janke & Kunkel, IKA-Werk) with speed control (Tissumizer® High Torque, Tekmar) setting at 75 before being homogenized in the Caframo apparatus.

Western immunobloting procedures for serum IgA were the same as for the pIgR and SC, except for the use of peroxidase-conjugated goat anti-mouse IgA antibody ( chain specific, Nordic Immunological Laboratories, Maidenhead, England). The PVDF membrane was incubated for 90 min in a 1:80,000 dilution of this reagent.

Bound peroxidase activity was revealed in the washed PVDF membranes by incubation in enhanced chemiluminescence reagent (Amersham Canada, Oakville) for 50-60 s. The membrane was then exposed in the dark to Kodak X-Omat K XK-1 film for 30 s to 1 min, and the intensity of bands was quantified by densitometer scanning (Dual-wavelength Flying-spot Scanner, Shimadzu, Kyoto, Japan). In relation to the study of the pIgR and biliary SC, the curve relating band intensity to the quantity of rat SC standard yielded an r² value always exceeding 0.98. Results pertaining to serum concentrations of IgA cross-reactive molecules were expressed in arbitrary units following scanning densitometry.

The rabbit antiserum to rat SC (gift of B. Underdown, McMaster University) was mainly of IgG class and exhibited no cross-reaction with rat serum according to Ouchterlony assay. The specificity of the antiserum was verified in the present investigation by means of electrophoresis and Western immunoblotting. These procedures were performed as described for the samples of hepatic, intestinal and biliary material, except that 70 g/L polyacrylamide gels were used and 1 µg of the antibody was absorbed for 1 h at room temperature with 4.3 µg of rat SC prior to use of the antibody as a first-stage immunoblotting reagent. A positive control was included in which a sample of antiserum was similarly incubated in PBS for 1 h. Samples of mouse liver, intestine and bile as well as rat SC and rat liver failed to yield a detectable reaction with the absorbed antiserum even if the PVDF membrane was exposed to X-ray film for as long as 5 min. An identical outcome was obtained with membranes exposed only to the second-stage reagent, i.e., peroxidase-conjugated goat anti-rabbit IgG. Finally, negative control samples prepared from mouse spleen, kidney and serum failed to exhibit reactivity with the anti-SC antiserum following exposure of PVDF membranes to X-ray film for as long as 5 min.

The specificity of the peroxidase-conjugated goat anti-mouse IgA was verified using preparations of purified mouse myeloma IgA and IgM (TEPC 15 and TEPC 183, respectively, Sigma Chemical) and of purified mouse IgG (Sigma Chemical) as described elsewhere (Ha et al. 1996). Moreover, the reagent yielded a reaction product when used immunohistochemically with intestinal sections rich in IgA-containing cells but failed to yield a reaction product with negative control tissue sections from mouse kidney (Ha et al. 1996).

RESULTS

Table 1. Body weights, food intakes, weight and protein concentrations of liver and intestine of mice when weaned or after 2 wk of free access either to the complete diet or to the low protein diet1 [View Table]

The cumulative food intake of the malnourished group (i.e., over the 14-d experimental period) was 43% of that exhibited by the well-nourished controls (Table 1). On a daily basis, however, this represented an intake of dietary energy and micronutrients comparable, on a body weight basis, to that of the well-nourished group (results not shown). This outcome is consistent with previous findings pertaining to the low protein system used in this investigation, viz., that serum triiodothyronine concentration remains comparable to that of well-nourished weanling mice (Filteau and Woodward 1987, Woods and Woodward 1991), whereas animals malnourished by means of restricted intake of a complete diet exhibit rapid and profound depression in the serum concentration of this hormone (Filteau and Woodward 1987). The wasting disease of the low protein group, therefore, resulted from dietary imbalance rather than from an insufficient intake of energy or of nutrients other than protein.

On the basis of comparison with the zero-time control group, both the liver and the intestine increased in weight in the well-nourished control group during the 14-d experimental period, although protein concentration was not affected in either organ in this group (Table 1). In the malnourished mice, both organs exhibited weight loss (in comparison with the zero-time control group) and a low protein concentration (relative to well-nourished controls).

A representative Western immunoblot of the intestinal pIgR is shown in Figure 2 together with lanes of rat SC standard. Mice of each group exhibited one or both of two bands with MW close to 97 and 116 kDa. Molecular heterogeneity has been demonstrated by Kuhn et al. (1983) in the pIgR of the mammary gland and the liver in rabbits and by Kloppel and Brown (1984) in the biliary SC of rats. Heterogeneity in the pIgR is therefore an established phenomenon, although it apparently has not been reported previously in connection with the intestine. The mammalian pIgR is heavily N-glycosylated (Piskurich et al. 1995), and heterogeneity in the molecule could derive from variations in this post-translational modification.

A representative Western immunoblot of biliary SC, including lanes of rat SC standard, is shown in Figure 3. Two immunoreactive bands were revealed in the bile and exhibited MW between 80 and 97 kDa (Fig. 3), a result similar to findings reported in a study of rats (Kloppel and Brown 1984). Bile from Group C mice also yielded a band (Fig. 3, lanes 6 and 9) at a MW of approximately 340 kDa. A band was also revealed in the same position by immunoblotting with peroxidase-conjugated anti-mouse IgA (results not shown). No such band was detected in the bile of Group LP (lanes 7 and 10) or Group B (lanes 5 and 8) mice even when the volumes of bile loaded per lane were 20- and 10-fold, respectively, the volume of bile loaded from Group C mice.

The concentrations of biliary SC and hepatic pIgR are shown in Figure 4. Group C mice exhibited 12.0- and 9.5-fold greater concentrations of biliary SC and hepatic pIgR, respectively, than the 19-d-old zero-time control group. The biliary SC and hepatic pIgR concentrations of LP group mice, however, were 50% and 21%, respectively, of the levels found in Group B mice and were only 4% and 2%, respectively, of the levels in Group C mice.

In contrast to the results pertaining to bile and liver, the intestinal pIgR concentration did not differ among the three groups of mice (Fig. 4). Because of this finding, a supplementary immunohistological study was conducted using methods described elsewhere (Ha et al. 1996). Dilutions of rabbit anti-rat SC and of peroxidase-conjugated goat anti-rabbit IgG (Bio-Rad) antibody were 1:1000 and 1:2000, respectively. Duodenal and jejunal sections exhibited an intense reaction product distributed throughout the epithelial cells in both Group C (n = 4) and Group LP (n = 2) mice (results not shown). These immunohistological results were therefore consistent with the outcome of the Western immunoblotting study of the intestinal crude particulate fraction. Duodenal sections did not reveal a reaction product when stained with only the peroxidase-conjugated second antibody. Spleen sections stained with both primary and secondary antibodies served as additional negative controls, and failed to exhibit peroxidase reaction product (result not shown).

Hepatic and intestinal pIgR (i.e., anti-rat SC immunoreactive material), expressed on a per organ basis, are shown in Figure 5. A significantly greater total amount of pIgR in both organs was found in Group C mice relative to the zero-time control group. In contrast, a profound reduction of the total amount of hepatic pIgR, to 9.6% of the quantity in Group B liver (and only 0.4% of the quantity in Group C), was found in Group LP mice. Although the concentration of intestinal pIgR was not different in Groups C and LP (Fig. 4), the total intestinal pIgR of Group C mice was about threefold higher than in Group LP animals because of the intestinal growth achieved by the former group (Table 1). Group LP mice maintained intestinal pIgR, on a per organ basis, relative to Group B mice (Fig. 5).

On the basis of the foregoing, the bands of MW ~150, 340 and >340 kDa were designated as monomeric (m), dimeric (d) and polymeric (p) IgA, respectively. The sum of the densitometer readings of the three bands in immunoblots of serum from Group LP mice (arbitrary units, mean = 125.7 × 10³) was 3.1-fold the value obtained for Group C mice (arbitrary units, mean = 37.1 × 10³), a finding that is consistent with a previous report that the low protein model produces an elevation in serum IgA concentration as is also common in malnourished humans (Ha et al. 1996). Each of these three presumptive forms of IgA was found in high concentration in the serum of Group LP mice compared with the serum of Group C animals (Fig. 7). This was particularly the case for the dimeric form, which was found at a concentration 7.6-fold that of Group C, whereas the monomeric and polymeric forms were found at concentrations that were 2.7- and 2.2-fold, respectively, the levels in Group C. 

DISCUSSION

Results from diverse experimental systems point to the quantity of pIgR as a determining factor in relation to the rate of transcytosis of IgA both in vitro (Kaetzel et al. 1991, Phillips et al. 1990) and in vivo (Prabhala and Wira 1991). The present results demonstrate reduction in the quantity of intestinal and hepatic pIgR in a model of wasting disease in which biliary IgA concentration and the quantity of IgA within the contents of the intestinal lumen both were low despite a high serum IgA concentration (Ha et al. 1996). The low mucosal concentrations of secretory IgA in this experimental system are therefore attributable substantially to a depressed expression of the pIgR. Others have proposed that PEM exerts a depressive influence on the transport of IgA to its mucosal sites of action (Chandra and Wadhwa 1993, McMurray et al. 1977). The present results are consistent with this proposition, and they initiate an understanding of the mechanism whereby PEM exerts a consistent depressive influence on the concentration of mucosal secretory IgA (Chandra 1991).

Low concentrations of free SC have been reported in tears of wasted children (Watson et al. 1985) and in tears, saliva and intestinal washings of weanling rats subjected to PEM by way of a low protein diet regimen (Sullivan et al. 1993). The low concentration of biliary SC associated with PEM in the present investigation is therefore consistent with previous information relating to this molecule in other external secretions in PEM. Functional inefficiency on the part of the hepatic pIgR in PEM may be inferred from the discovery of free SC, albeit in low concentration, in the bile of Group LP mice (as in Group C mice) in the apparent absence of secretory IgA which was detectable only in the well-nourished controls. The mechanism whereby the low protein protocol may impair the function of the pIgR remains to be investigated. In this regard, however, the G protein subunit, Gs, stimulates transcytosis of the pIgR by way of cAMP and protein kinase A (Hansen and Casanova 1994), whereas desensitization of the protein kinase A response to cAMP was reported in the liver of weanling rats subjected to wasting PEM (Rozwadowski et al. 1995). In addition, disruption of microtubules impairs transcytosis of IgA (Breitfeld et al. 1990), and such a disturbance in cellular structure seems probable in diverse tissues, e.g., the cerebral cortex (de Mattos et al. 1994) and the thymic epithelium (Mittal and Woodward 1986), in PEM. Alternatively, production of dIgA lacking the J chain peptide would also reduce the efficiency of pIgR-mediated transcytosis of IgA by the liver (Hendrickson et al. 1995). Low mucosal IgA concentrations such as occur in the present system of experimental PEM (Ha et al. 1996) may therefore result both from reduced expression of the pIgR and from functional inefficiency on the part of the small quantity of receptor that is expressed.

In the present investigation, the intestine and liver responded differently to PEM in terms of the cellular pIgR concentration. This outcome is easily reconciled with existing information demonstrating that the neuroendocrine regulation of pIgR synthesis is site specific (Lambert et al. 1994). Moreover, the hepatic and intestinal pIgR exhibited different developmental kinetics in the present experimental system. The hepatic pIgR increased 9.5-fold in concentration in well-nourished mice between 19 and 33 d of age. In contrast, the intestinal pIgR concentration did not differ between zero-time control and well-nourished mice. The present study therefore seems to have been conducted at a stage of murine development at which the hepatic pIgR is more physiologically labile than the intestinal pIgR. A clear understanding is lacking as to the neuroendocrine mechanisms whereby site-specific control of pIgR synthesis is achieved.

The magnitude of PEM-associated depression in hepatic pIgR concentration in the present experimental system implicates an influence on the synthesis of this protein. Depression of hepatic protein synthesis, however, is not a general phenomenon in PEM (Straus et al. 1994). In a model of wasting malnutrition involving weanling rats fed a low protein diet, the hepatic messenger RNA levels for some proteins (albumin, transthyretin, carbamyl phosphate synthetase and alcohol dehydrogenase) were low relative to levels in well-nourished controls, whereas the messenger RNA levels for several other proteins (hypoxanthine-guanine phosphoribosyl transferase, ubiquitin, H-ferritin and insulin-like growth factor binding protein-4) were either unaffected or high (Straus et al. 1994). The pIgR therefore seems likely to be among those hepatic proteins that are most sensitive to PEM.

The distribution of molecular forms of serum IgA in PEM has not been reported prior to this investigation, and was studied as an index related to the quantity of functional hepatic pIgR. The present results pertaining to well-nourished control animals (59% mIgA and 41% in the combined dimeric and polymeric IgA fraction) are comparable, quantitatively, to those reported by Delacroix et al. (1985) in relation to adult mice. Presuming that PEM does not increase the rate of synthesis of IgA, the high serum level of this Ig (all forms) in the LP group focuses attention on mechanisms whereby this molecule is removed from the blood. The main point pertains to the disproportionately high concentration of dIgA found in the serum of the LP group. In mice, mIgA is removed from the blood plasma by a hepatic asialoglycoprotein receptor (Moldoveaunu et al. 1988), whereas dIgA is thought to be removed mainly by way of pIgR-mediated hepatobiliary transport (Kerr 1990). The abundance of dIgA relative to mIgA in the blood of the Group LP mice is therefore easily reconciled with the low expression of the hepatic pIgR in this group. Like dIgA, higher MW forms of this Ig are also thought to be removed from the blood by the hepatic pIgR (Kerr 1990). The mechanism whereby the low protein protocol induced an overabundance of dIgA relative to pIgA in the blood is therefore not clear. Song et al. (1995), however, showed that conditions that restrict the number of cellular IgA receptor sites will limit the pIgR-mediated transcytosis of dIgA more severely than of pIgA (tetrameric form). It is unlikely that a shift toward dIgA in the blood is significant, of itself, to the immunopathology of PEM. This finding, however, is consistent with the profound decrease in hepatic pIgR in the LP group and thus provides independent support for the conclusion that the effect of the low protein protocol on the hepatic pIgR bears functional importance.

In summary, we investigated a murine model of weanling PEM that mimics the human condition in terms of important aspects of IgA concentration (Ha et al. 1996, this study) and systemic acquired immunity (Woods and Woodward 1991). Use of this experimental system highlights the resistance of the systemic and mucosal humoral effector compartments to the influence of wasting PEM (Ha et al. 1996) but, at the same time, provides evidence that the unique mucosal requirement for epithelial transport of IgA is sensitive to PEM (this investigation). In PEM, therefore, expression of functionally efficient pIgR may limit IgA concentrations in external secretions independently of any influence on IgA-containing cell numbers (Fig. 8). The result is a deficiency of the blocking antibody action of IgA that counteracts the propensity of disease-causing microorganisms to adhere to, and sometimes to penetrate, mucosal epithelia. In addition, an important implication of these results relates to the therapeutic enhancement of mucosal immunity in wasted subjects. An intervention directed only toward increasing IgA-containing cell numbers would be unlikely to enhance mucosal humoral immunity in PEM, whereas stimulation of pIgR synthesis and function might promote a meaningful increase in the mucosal IgA concentration. Polymeric immunoglobulin receptor synthesis is regulated by diverse hormonal (including cytokine) and neural stimuli (Lambert et al. 1994, Phillips et al. 1990), but its responsiveness to such stimuli in wasted subjects remains to be determined.