Impairment of Mucosal Immunity by Parenteral Nutrition: Depressed Nasotracheal Influenza-Specific Secretory IgA Levels and Transport in Parenterally Fed Mice

Abstract

Objective

To examine the effects on mucosal selective transport of polymeric IgA (pIgA) and the ability of exogenous pIgA to provide protection despite altered mucosal transport.

Summary Background Data

Parenteral nutrition significantly impairs established antipseudomonal immunity and IgA-mediated antiviral immunity in association with gut-associated lymphoid tissue mass atrophy. Lack of enteral feeding also induces mucosal effects.

Methods

After immunization, nasotracheal levels of influenza-specific IgA were measured in cannulated mice randomized to chow feeding or parenteral nutrition. Nonimmune animals were randomized to chow or total parenteral nutrition, and after 5 days of diet were given a mixture of two antiinfluenza monoclonal antibodies, pIgA and IgG. Four hours after injection, nasal washes were collected and influenza-specific antibody levels were determined by enzyme-linked immunosorbent assay to calculate the selective transport index of IgA relative to IgG. In the final experiment, immunized animals were randomized to chow or parenteral feeding, and after 5 days, parenterally fed animals received either normal mouse serum or antiviral pIgA before viral challenge. Viral shedding was measured at 42 hours after challenge.

Results

Parenteral nutrition significantly reduced virus-specific IgA in nasotracheal washes. Parenteral nutrition depressed the selective transport index, demonstrating impaired mucosal transport of pIgA. Parenterally fed animals given specific antiviral pIgA but not normal mouse serum eliminated virus from the airway and regained mucosal protection, demonstrating adequate residual transport for immunity if adequate pIgA is present.

Conclusion

Although both decreased IgA production due to gut-associated lymphoid tissue atrophy and impaired mucosal transport occur when enteral feeding is not provided, residual transport can provide antiviral protection if exogenous antiviral pIgA is available. Production, rather than transport, may be the most important factor in maintaining established respiratory tract IgA-mediated immunity.

Our work has implicated diet-induced changes in mucosal defenses as a factor in the increased incidence of pneumonia seen with parenteral feeding. ^1,2 These defenses include nonspecific immune defenses such as lactoferrin, peroxidases, defensins, and other inhibitory substances of bacterial growth as well as specific defenses. ³ Specific immunity is primarily due to IgA, which is produced by immune cells within the lamina propria and transported across the mucosal epithelia by secretory component (SC). ⁴ B and T cells producing and controlling IgA production are sensitized to luminal antigens within the Peyer’s patches of the gut-associated lymphoid tissue (GALT), although an upper respiratory site, the nasal-associated lymphoid tissue, appears active in some animal species. ⁵ The sensitized cells ultimately home to the lamina propria of both intestinal and extraintestinal sites (e.g., respiratory tract), where they produce antigen-specific IgA. IgA is transported into the mucosal secretions to coat and protect moist mucosal surfaces by neutralizing or otherwise preventing attachment and infection by viruses and bacteria. ^5–8

Our work has demonstrated significant defects in specific mucosal immunity during parenteral nutrition (TPN). ⁹ TPN downregulates the entire system by reducing numbers of both B and T cells within all GALT compartments and lowering secretory IgA levels throughout the mucosal immune system. ^9,10 Reduction of the intestinal Th2-type IgA-stimulating cytokines, interleukin 4 (IL-4) and interleukin 10, with intravenous TPN correlates with reduced luminal IgA levels. ¹¹ These changes may explain the increase in intestinal bacterial translocation and the loss of established respiratory antiviral ¹² and antibacterial ¹³ defenses in TPN-fed animals.

Although it is tempting to attribute the lowered respiratory and gastrointestinal mucosal secretory IgA levels solely to TPN-induced loss of GALT cell mass, impaired transport by the epithelial cell itself is possible, because parenteral feeding produces mucosal atrophy and changes in cytokine production that could alter epithelial transport function. ¹⁴ This study investigates the mucosal transport of intravenously administered polymeric IgA (pIgA) in parenterally fed mice using the selective transport index (STI) and the ability of intravenously administered influenza-specific monoclonal pIgA to reverse TPN-induced impairments in antiinfluenza respiratory mucosal immunity.

METHODS

Animals

Six- to 8-week old male Institute of Cancer Research mice (Harlan, Indianapolis, IN) were housed in a conventional facility accredited by the American Association for the Accreditation of Laboratory Animal Care under controlled conditions of temperature and humidity with a 12/12-hour light/dark cycle. Before the experiment, mice were given free access to water and commercial chow. During the experiments, the mice were housed in metal, wire-grid—bottom metabolism cages to eliminate coprophagia and the ingestion of bedding. All protocols were approved by the institutional animal care and use committee.

Experimental Protocol

Experiment 1: Nasal Specific Antibody Determination

Mice were immunized intranasally while awake with 20 μL phosphate-buffered saline (pH 7.4) containing 100 mouse 50% mouse infectious dose (MID₅₀) of A/PR8 (H1N1) influenza virus. Three weeks later, the mice received venous catheters as previously described. ^9–13 All animals were allowed free access to chow for 2 days, then randomized to chow or intravenous TPN for an additional 5 days. The chow group received chow, intravenous saline, and free access to water. The TPN group received a standard TPN solution (4.1% amino acids, 34.3% glucose, electrolytes, and multivitamins with a nonprotein calorie-to-N ratio of 740 kJ/g N), which was increased during the next 24 hours to 10 mL/day to meet the energy and protein requirements of mice (1,619 kJ/kg per day of nonprotein calories and 14 g protein/kg per day). ¹⁵ After 5 days of chow (n = 15) or TPN (n = 17), the animals were anesthetized and exsanguinated.

Nasal washes were collected through a midline incision made over the ventral aspect of the trachea slightly anterior to the thoracic inlet. The trachea was clamped off at the thoracic inlet and 400 μL phosphate-buffered saline was slowly injected into the tracheal lumen cephalad to the obstruction using a 25-gauge B-bevel needle attached to a tuberculin syringe. Nasal washes for enzyme-linked immunosorbent assay (ELISA) analysis were placed on ice immediately on collection and stored at 4°C until assay.

Experiment 2: Selective Transport Index Determination

Nonimmune mice were surgically instrumented with intravenous catheters, given access to chow for 2 days, and then randomized to TPN (n = 9) or chow (n = 11). After 5 days on their respective diets, mice were injected intravenously through their indwelling catheters with 400 μL ascites fluid containing 1,700 μg pIgA and 1,460 μg IgG₁ antiinfluenza monoclonal antibodies. Before intravenous injection, all ascites samples were centrifuged for 5 minutes to preclude the formation of emboli. Nasotracheal lavages were performed as above, and the resulting washings were assayed for influenza-specific antibodies by ELISA. Blood was collected by cardiac puncture and the serum was retained for antibody determination. Monoclonal antibody content of the samples was determined by ELISA, and the transport ratio between the two monoclonals in a single animal, represented by the selective transport index (STI), was calculated as

An STI of 1 indicates that both pIgA and IgG leaked or were transudated into nasal washes at the same rate, while an STI of greater than 1 implies selective transport of pIgA relative to IgG₁. ¹⁶

Experiment 3: Restoration of Protection by Passively Administered Antibody

Mice were immunized intranasally as above. Three weeks later, immune mice were surgically instrumented for TPN or saline administration and nonimmune mice for saline administration only. Mice were allowed to recover for 2 days and then randomized to one of three groups: chow-fed immune (n = 5), chow-fed nonimmune (n = 5), or TPN-fed immune (n = 8). After 5 days of diet, the TPN-fed immune group was further subdivided into two groups: TPN-fed and injected with pIgA (n = 4) or TPN-fed and injected with normal mouse serum (n=4). The TPN-fed pIgA-injected group received 200 μL intravenously of the influenza-specific monoclonal pIgA-containing ascites previously shown to protect the noses of nonimmune mice from influenza infection, ¹⁷ whereas the TPN-fed normal mouse serum-injected group received 200 μL normal mouse serum nonreactive with influenza virus. The immune chow-fed and a nonimmune, noncannulated control group also received normal mouse serum. Four hours after injection of normal mouse serum or influenza-specific monoclonal pIgA, all mice were challenged intranasally while awake with 100 MID₅₀A/PR8 (H1N1) influenza virus in 20 μL phosphate-buffered saline. Forty-two hours later, mice were killed, the trachea was clamped at the thoracic inlet, and 500 μL phosphate-buffered saline was injected cephalad using a 25-gauge needle on a tuberculin syringe. Nasal washes were placed on ice and immediately assayed for virus. Details of virus preparation, viral assay, and ELISA assay of nasotracheal influenza-specific IgA levels have been previously described. ^12,17

Monoclonal Antibodies

Hybridomas producing monoclonal antibody H37-66 (pIgA or monoclonal pIgA) and 2R6 (IgG₁), directed toward the H1 hemagglutinin of PR8 influenza virus, were the gift of Dr. Walter Gerhard (Wistar Institute, Philadelphia) and were generated as reported. ¹⁸ Hybridomas were grown in BALB/c mice as previously described, ¹⁷ and the ascites fluids were pooled and frozen. The antibody content of the pools was determined using a radioimmunoassay kit according to label instructions (ICN ImmunoBiologicals, Lisle, IL). The pooled ascites fluids used contained approximately 8,500 μg influenza-specific monoclonal pIgA per milliliter and 7,300 μg influenza-specific IgG₁ per milliliter.

Statistics

All data are expressed as the mean ± SE. Nonparametric (Mann-Whitney) and parametric (unpaired two-tailed Student t test, analysis of variance, and Fisher exact test) analyses were carried out on a Macintosh Performa using Statview 4.2 (Brain Power, Inc., Calabasas, CA) software.

RESULTS

Experiment 1

Levels of influenza-specific secretory IgA antibody in undiluted nasal lavages were expressed as OD_405nm (Fig. 1). Nasal secretions of chow-fed immune animals had a mean OD of 0.393; those of TPN-fed mice had a mean OD of 0.253. The nasal influenza-specific secretory IgA of the TPN-fed group was significantly lower than that of chow animals (P < .005).

Figure 1. Chow-fed mice had significantly higher amounts of IgA in nasal washes than parenterally fed mice.

Experiment 2

There were no significant differences in pre- and postexperiment body weight in each group or between chow-fed and TPN-fed mice. The mean STI of chow-fed mice was 4,633 ± 297, reflecting selective transport of IgA into nasal secretions. The STI of TPN-fed mice dropped significantly to 1,164 ± 859 (P < .005) (Fig. 2).

Figure 2. Parenteral feeding significantly reduced nasal polymeric IgA selective transport compared with chow-fed mice (P < .005).

Experiment 3

As positive and negative controls, all five chow-fed immune control mice were protected against viral infection, whereas all five nonimmune chow-fed mice shed virus into their nasal secretions (P < .01). All five TPN-fed immune mice given placebo normal mouse serum had complete loss of effective immunity and shed virus at 42 hours. However, all four TPN-fed immune mice given antiviral pIgA were protected (P < .05 vs. placebo-injected TPN-fed mice), establishing that residual transport, although depressed in parenterally fed animals, was sufficient to restore secretory IgA-mediated protection against influenza virus when pIgA was provided exogenously.

DISCUSSION

In 1971 Heremans and Bazin ¹⁶ reported that IgA plasma cells producing antibody specific for orally administered antigens were detected not only in the gut of immunized animals, as expected, but also in extraintestinal lymphoid organs such as mesenteric lymph nodes and spleen. Since the original observation, the results have been confirmed in several extraintestinal sites: gnotobiotic rats fed killed Streptococcus mutans had specific IgA antibody in both saliva and milk, ¹⁹ and human volunteers ingesting killed S. mutans had specific antibody of the IgA class in both saliva and tears. ^20,21 Mice fed protein antigens had IgA-producing cells in the gut and in secretory glands. ²¹ These and many other observations led to the concept of a common mucosal immune system. ^20,22

We have shown that the feeding of mice parenterally rather than enterally has deleterious effects on the common mucosal immune system: GALT atrophy occurs with associated decreases in Peyer’s patches cells, lamina propria lymphocytes, and intraepithelial lymphocytes;⁹ total secretory IgA levels decrease in the gastrointestinal ⁹ and respiratory ¹⁰ tracts; and functional mucosal immunity is impaired in both the upper ^10,12,13 and lower respiratory tracts. ^13,23 The depressed specific secretory IgA could result from lowered production of pIgA by the lamina propria plasma cells, resulting from either a defect in the plasma cells themselves or to a decrease in their numbers; from a defect in the SC-mediated pIgA transport system; or from a combination of transport and production defects.

Under normal conditions, pIgA antibody produced by plasma cells in the lamina propria binds to SC, the pIg receptor, on the basolateral surface of the mucosal epithelial cell and is transported through the cell to its apical surface, where the antibody and an accompanying portion of the pIg receptor are cleaved off the cell to become secretory IgA. ²⁴ Unfortunately, no antibody specific for murine SC is available to measure this transport mechanism directly in vivo. Hence, the selective transport concept was developed, ¹⁷ comparing the transport of an antibody requiring SC for transport (pIgA) with that of an antibody (IgG₁) to which normal epithelium is relatively impermeable. Our results demonstrate that this selective transport of pIgA by SC is depressed but not absent in TPN-fed mice (see Fig. 1), suggesting that SC is produced at a reduced level in these animals and demonstrating that lack of enteral feedings reduces this specific epithelial function. The respiratory mucosa itself remains intact in TPN-fed mice because there was no evidence of leakage of passively administered IgG₁ into nasal secretions.

The mechanism for reduced SC can be at least partially explained by changes in cytokines induced by TPN. SC synthesis is known to be under cytokine control: SC expression on human epithelial cells is regulated in a synergistic fashion by interferon-gamma (IFN-γ) and IL-4. ²⁵ We recently showed that although the level of intestinal IFN-γ remains unchanged by the diet, the level of IL-4 in the intestinal tracts of TPN-fed mice decreases significantly. This decrease correlates with a decline in intestinal IgA levels. ¹¹ We hypothesize that the synergism between IL-4 and IFN-γ is lacking in TPN-fed animals, resulting in a lowered level of SC synthesis. In enterally nourished chow-fed mice, normal interaction occurs between IL-4 and IFN-γ, allowing normal SC expression.

Although the SC-mediated mucosal epithelial antibody transport system appears compromised in TPN-fed mice, enough function remains for intravenous administration of influenza-specific monoclonal pIgA antibody to provide passive protection of the nose against influenza infection (see Fig. 2). This protection is IgA mediated ⁸ and requires the selective transport of pIgA through the mucosal epithelial cells. ¹⁷ Because the upper respiratory tract SC-mediated transport system can transport pIgA and protect passively immunized TPN-fed mice when an adequate level of influenza-specific pIgA is present, the defect underlying the loss of respiratory immunity in TPN-fed mice appears more related to inadequate production of pIgA in the lamina propria as a result of reduced GALT cell mass and loss of Th2-type IgA-stimulating cytokines than to an inadequate ability to transport the IgA once it has been produced by plasma cells. This putative antibody shortage could be due to an insufficient number of antibody-producing plasma cells, to an intrinsic defect in the plasma cells themselves, or both.

In conclusion, severely impaired mucosal immunity is observed in TPN-fed mice. This immunologic impairment involves atrophy of the GALT, a reduction in intestinal and respiratory IgA, and loss of respiratory immunity, as evidenced by the loss of protection against influenza virus in TPN-fed immune mice. The upper respiratory tract defect appears to encompass deficits in both antibody transport and antibody production in the mucosa of the nasotracheal region, but adequate transport is preserved to provide protection in the presence of adequate IgA.