IgA Antibodies to Dietary Antigens and Lectin-Binding IgA in Sera From Italian, Australian, and Japanese IgA Nephropathy Patients

IgA Antibodies to Dietary Antigens and Lectin-Binding IgA in Sera From Italian, Australian, and Japanese IgA Nephropathy Patients Rosanna Coppo, MD, Alessandro Amore, MD, Dario Roccatelio, MD, Bruno Gianoglio, MD, Andrea Molino, MD, Giuseppe Piccoli, MD, Anthony R. Clarkson, MD, Andrew J. Woodroffe, MD, Hideto Sakai, MD, and Yasuhiko lomino, MD • We studied serum IgA as antibodies to dietary antigens (Ag), as lectin-binding molecules, and as conglutininbinding immune complexes (lgAIC) in people from geographical areas in which IgA nephropathy (lgAGN) is particularly frequent. Sera from 63 Italian, 21 Australian, and 25 Japanese patients affected by IgAGN and 24 Italian, 20 Australian, and 40 Japanese healthy controls were studied. Increased values of IgAIC were detected in 42.8% of Italian patients, while only in 23.8% and 8% of Australian and Japanese patients, respectively. Mean values were significantly increased only in Italian patients (P<0.0001). Positive values of IgA antibodies against dietary Ag had variable prevalences, but again Italian patients showed the highest frequency, from 190/0 to 28.5% versus 0 to 38% in Australians and 0 to 16% in Japanese. Mean values of these antibodies were not significantly increased in any patient groups in comparison to the corresponding healthy populations. However, patients with elevated values of IgAIC had significantly higher serum concentrations of antibodies to alimentary components and a linear correlation was found between IgAIC and some IgA antibodies to food components. The relationship between these two series of data was particularly evident for Italian and Australian IgAGN patients. Moreover, the patients with positive data tended to have a cluster of increased levels of IgA antibodies against several alimentary Ag at the same time. A linear correlation was evident between values of IgA antibodies to gluten fractions and to heterologous albumins. None of these correlations was evident among healthy controls. Besides the role of the antibodies, some IgA in sera from the patients investigated were able to bind severallectins by a nonimmune bond, which was inhibited by competitive sugars. Comparing patients' data with the corresponding controls, we found increased lectin-binding IgA activity in up to 72% of Japanese patients, in up to 56% of Australians, and in 6% to 33% of Italians. Only in Japanese patients' sera were mean values of lectin-binding IgA activity significantly increased, but without correlation with IgAIC levels. These data suggest that serum IgA in IgAGN patients may participate in immune complex formation with IgA binding to alimentary antigens and also forming nonimmune complexes by IgA-lectin interactions. © 1991 by the National Kidney Foundation, Inc. INDEX WORDS: IgA nephropathy; antibodies to dietary antigens; lectins; lectin-binding IgA; IgA immune complexes.

I

N RECENT YEARS, an area of research in IgA nephropathy (IgAGN) has been devoted to identifying the antigen(s) (Ag) eliciting the IgA immune response. In favorable conditions, this leads to electron-dense_ mesangial deposits containing IgA and consequent renal tissue damage. 1-5 Several experimental models demonstrated that, besides infectious agents,6 common alimentary Ag can be immunogens, inducing a mucosal immune system reaction producing specific IgA antibodies and IgA immune complexes (IgAIC) eventually delivered to the mesangium. 7.8

From the Institute of Nephro-Urology. University of Turin. Turin. Italy; Renal Unit. Royal Adelaide Hospital. Adelaide. South Australia; and the Department of Internal Medicine. Tokai University. Isehara City. Japan. Address reprint requests to Rosanna Coppo, MD, Divisione di Nefrologia e Dialisi, Ospedale Infantile Regina Margherita, Piazza Polonia 94, 10126 Torino, Italy. © 1991 by the National Kidney Foundation, Inc. 0272-6386/91//704-0023$3.00/0 480

In IgAGN patients, sporadically increased levels of serum IgA antibodies to alimentary Ag have been detected. 9-11 IgA antibodies to gliadin Ag seem to be particularly frequent, even though the reported prevalences in IgAGN patients varied from 0 to 75%,1115 possibly due to the different techniques employed with variable sensitivity for borderline values. Other researchers sporadically found various alimentary antigens in IgAIC.16.17 Similar inconsistent data were reported by analyzing the antigenic component of mesangial deposits. 9.18 It has been recently demonstrated that macromolecular IgA can be formed by the reaction between a given Ag and the specific IgA antibody, as well as between IgA and lectin-like proteins, with formation of nonimmune IgA complexes. 19 Lectins are carbohydrate-binding proteins of nonimmune origin.20 They have a heterogeneous derivation from bacterial, viral, plant, or animal products. They can bind oligosaccharides contained in immunoglobulins, particularly IgAl,

American Journal of Kidney Diseases, Vol XVII, No 4 (April), 1991: pp 480-487

481

IGA TO DIETARY AG AND LECTIN-BINDING IGA

which has a set of galactose/ N-acetyl-galactosamine residues in the hinge region 21 and polymeric IgA rich in terminal galactosyl residues. 22 Lectin-binding sites exist in several glomerular structures 23 and gliadin, the lectin-like fraction of gluten, can bind to mesangial cells in culture. 24 Gliadin can also bind purified IgA 3 and increase the metabolic activation and respiratory burst of neutrophils. 25 Intra-aortically injected IgA-Iectin nonimmune complexes can induce IgA and C3 mesangial deposits. 26 A decrease in glomerular sialic acid lectin-binding sites has been observed in patients wth IgAGNY Lectins can also bind to carbohydrate residues on the intestinal epithelial surface, increasing the permeability to alimentary antigens in the gut lumen. 28·30 Of interest is the observation that alimentary lectins, such as gluten and soya, are abundantly eaten by populations with a high prevalence of IgAGN, such as in Mediterranean Europe and the Far East. We have demonstrated that oral immunization with gliadin can experimentally induce IgA mesangial deposits. 31 From these data, one may hypothesize that IgA complexes may be formed by an IgA antibody response to alimentary Ag with IgAIC formation, as well as by a nonimmune interaction between IgA and environmental lectins. IgAGN is particularly prevalent in Mediterranean Europe (ie, France, Italy, Spain), Japan, and Australia, while it is far less common in North America and Northern Europe. 32 .33 Even though these prevalences may be strongly influenced by different biopsy policies, 34 a variable frequently of IgAGN in diverse countries remains indisputable. Genetic predisposition or environmental factors may favor this dishomogenous geographic distribution. The aim of this study was to investigate sera of IgAGN patients from geographical areas in which this nephropathy is particularly prevalent, namely Italy, Australia, and Japan, and to evaluate levels of IgA antibodies to dietary Ags, lectin-binding IgA activity, and IgAIC. PATIENTS AND METHODS

Patients and Controls We studied 109 patients with primary IgAGN, of whom 63 were Italians (54 males, nine females, aged 15 to 65 years [mean, 35.5 years]), 21 Australians (11 males, 21 females aged 19 to 66 years [mean, 32.5 years]), and 25 Japanese (17 males, 8 females aged 16 to 52 years [mean, 33.6 years]).

As controls, we studied 84 healthy subjects, age- and sexmatched, of whom 24 were Italian, 20 Australian, and 40 Japanese.

Detection of IgA Antibodies to Dietary Antigens Antigens investigated were ovalbumin (OVA), bovine serum albumin (BSA) , the gluten fractions ethanol-soluble gliadin (glia-eth) and bicarbonate-buffer soluble gliadin (glia-BB), glutenin (glu), and the lectin fraction termed glyc-gli. Antigens. OVA in crystalline form was purchased from Calbiochem, La Jolla, CA, BSA, crude gluten, and gliadin from Sigma, St Louis, MO. The gluten fraction glutenin was prepared by stirring crude gluten for 12 hours at 22 °C in a low molarity acetic acid solution (0.05 moUL) and collecting the supernatant fluid after centrifugation at 2,500 rpm for 30 minutes. The gluten fraction glyc-gli was prepared as described by others2' Briefly, 25 g of gluten was stirred for 12 hours at 22 °C in 800 mL of 0.1 moUL acetic acid in 2.0 mol/L ethanol, pH 6.5. The supernatant collected after centrifugation at 2,500 x g for 15 minutes was added to 125 mL of2.2 moUL trichloroacetic acid and incubated in an ice bath for 1 hour. The precipitate, obtained by centrifugation at 5,000 x g for 1 hour at 4°C was redissolved in acetic acid, 0.1 moUL, and dialyzed for 2 days against phosphate-buffered saline (PBS), 0.15 moUL, pH 7.5. The soluble fraction was cleared by centrifugation at 5,000 x g for 30 minutes. Enzyme-linked immunosorbent assay (ELISA) for /gA antibodies to dietary antigens. Microplate wells (Nunclon, Nunc, Kanstrup, Denmark) were coated with gliadin dissolved in 70% ethanol, incubated at 22°C until ethanol evaporated, or with gliadin, glu, glyc-gli, OVA, or BSA in 0.1 moUL bicarbonate buffer, pH 9.6. For coating microplate wells, gluten-derived Ag were diluted to 50 jtg/mL and albumins to 10 jtg/mL. Plates were then incubated for one hour at 37°C and for 3 days at 4°C. After two washes with PBS 0.15 moUL, pH 7.2, containing 0.05% Tween 20 (PBS-Tween), the nonspecific binding was blocked with 0.1 % human serum albumin, 0.1 % gelatin in PBS for 2 hours at 22°C. After two washes in PBS-Tween sera previously diluted 1:50 in PBS-Tween containing l-a-methylD-glucopyranoside 1 moUL were incubated overnight at 4 °C in the Ag-coated wells. Plates were then washed and incubated with anti-IgA monoclonal antibody (TEC-IgA, Technogenetics, Turin, Italy) 1:500, 1 hour at 37°C and 30 minutes at 4°C. After washes, alkaline phosphatase-conjugated goat anti-mouse IgG-F(ab), antibodies (Technogenetics, Turin, Italy), diluted 1:500 were incubated for 1 hour at 37°C and 30 minutes at 4°C. After two additional washes, 4-p-nitrophenylphosphate 1.5 mg/mL in 0.1 moUL bicarbonate buffer, pH 9.6, was added to each well as specific substrate for alkaline phosphatase. Adsorbance at 405 nm was read on a multichannel spectrophotometer at the end of the linear phase of the reaction by an automated microplate photometer Dymatech MR 600 coupled to a microcomputer (Apple Computer, Sunnyvale, CAl. Values exceeding the upper 90% confidence limit in healthy controls from the corresponding nationality group were considered positive.

Detection of Circulating IgAIC The conglutinin solid-phase assay has been described in detail elsewhere. 35 Briefly, conglutinin purified from bovine sera was used for coating microplate wells (Nunclon, Nunc, Kanstrup, Denmark) in 0.1 moUL bicarbonate buffer, pH 9.6. Af-

482 ter incubation for 3 hours at 37°C and 3 days at 4 °C, the plates were washed with PBS-Tween. Sera, obtained after fasting overnight and previously diluted 1:3 in PBS-Tween were incubated 1 hour at 37°C and 30 minutes a t 4 C. ° The plates were then washed again and the revealing system, an anti-human IgA antibody conjugate with alkaline phosphatase (Sigma) diluted 1:2,000, was added. Results were expressed in optical density (OD) units. Intraassay data variations were corrected 3 • by calculating the results according to the formula : (P/Pt)x, where P was the mean OD value obtained by testing 30 healthy controls (0.25 OD); Pt was the normal value of a pool of healthy controls obtained in the daily protocol, and x the OD of duplicate test specimens. Values exceeding the upper 90% confidence limit in the healthy controls from the corresponding nationality group were considered positive.

Measurement of Serum /gA Capability of Binding Lectins (Lectin-Binding /gA Activity) The following lectins and inhibitory sugars were used for the assay : Phytolacca Americana (Pokeweed; Sigma) inhibited by l-a-methyl-D-glucopyranoside (l-mGIc) 1 mol/L, Glicine Maximum (Soybean; Sigma) inhibited by galactosamine 0.5 mol/L (Gal), Soya flour (Kl snc , Laboratorio Artigianale, Thrin, Italy) inhibited by I-mGIc 1 mollL, Gliadin (Sigma) inhibited by I-mGIc 1 mollL and N-acetyl-glucosamine 0.5 moll L (GIcNAc). The lectin fractions were obtained from soya flour and gliadin by 70% ethanol extraction. ELISA was used for measurement of the binding of IgA to lectins. The technique was similar to that described above for the detection of IgA to dietary antigens . Microplate wells were coated with lectins, 10 /tg/mL in carbonate buffer 0.1 mollL, pH 9.6 or in 70% ethanol , incubated at 22 °C until ethanol evaporated. Plates were washed with PBS-Tween and blocked with serum albumin and gelatin . Each serum was separately diluted 1: 50 in PBS-Tween and in each inhibitor sugar. Sera from each patient without and with the inhibitor sugar were tested, in duplicate, in consecutive wells. Incubation of sera, well washes, and incubation of the anti-IgA monoclonal antibody and of anti-mouse IgG-F(ab), alkaline-conjugated antisera were performed as described above for the detection of IgA to dietary Ag, albeit the antibodies were diluted either in PBS or in the inhibitor sugar solutions. The final OD at 405 nm obtained in the wells with the inhibitory sugar was subtracted from the value read in the wells without it. The lectin-binding activity of IgA was calculated as a ratio of this difference to the OD value obtained without inhibitory sugar. This ratio approaches 1as the capability of IgA to bind lectins increases. Values exceeding the upper 90 % confidence limit in healthy controls from the corresponding nationality group were considered positive.

COPPO ET AL

RESULTS

Levels of /gA Antibodies to Dietary Antigens Mean values of IgA to dietary Ag in IgAGN patients from Italy, Australia, and Japan were not significantly increased in comparison to the corresponding healthy control groups (Fig 1). The prevalence of positive data for IgA to dietary Ag ranged from 19 % to 42.8 % among Italians, from 0 to 38% in Australians, and from 0 to 16% in Japanese patients. By comparing the data of the three patient groups, no substantial differences were observed, but an increase was found in IgA to glia-BB and gJyc-gli in sera from Italian IgAGN patients versus Australian (Pnti-

po),li ... d . ' .

0

'"

o

(ITALY)

i

75

i

I

,

0 0

!

II

"l'

I

0

I

OVA

OVA

_

OS...

os.

GlI. 00

Gli. £lh

Gli.

Glle

B8

Clh

(

I

~ Iut

Statistics Statistical calculations were performed on a microcomputer (MacIntosh Plus, Apple Computer, Cupertino, CA) using a standard statistical package (Statview 512 , Brainpower, Calabasas, CA) . For comparison between patient and control groups, the nonparametric Mann-Whitney U test was used. The regression coefficient r was used to determine the correlation between the two series of data . The chi-square test was used to compare prevalences of positive data.

I ...

S

0 50

"

I. ,.. '.

( t) (~; ( t, I' fI

... ,. ..... 1. n c .o1

Cl ut

Glu t

~. :

J,

GI'oIc - G1I

Clyc - Cli

Glye - GII

IgAGN po tle n t.

~ Co n t r ol s

I stend erd de vielion

Fig 1. IgA antibodies to dietary antigens in IgAGN patients from Italy, Australia, and Japan and in corresponding healthy control groups. Shaded areas in the circles indicate the prevalence of positive data in the corresponding control population. Columns indicate mean values and bars 1 SO.

483

IGA TO DIETARY AG AND LECTIN-BINDING IGA

gens in Japanese controls were significantly increased in comparison to Australians (OVA, P
Mean levels of IgAIC were not significantly different among Italian, Australian, and Japanese IgAGN patients (Fig 2). However, by comparing patient data with the corresponding control group, a significant difference was found only for Italians (P
We considered two groups ofltalian patients, 27 with positive IgAIC data (>90th percentile) and 36 with normal values. Patients with elevated ti!.revalence of 1'03111Ve dllta

0.90 0.60

Correlations Among IgA Antibodies to Various Dietary Antigens

040 0.30

"'

0.20

0

" Q Q

ters of IgAIC had a significantly higher mean concentration ofIgA to gluten fractions and to heterologous dietary albumins than patients with normal values of IgAIC (Table 1). We failed to show a linear correlation between IgAIC and IgA to dietary Ag levels. Among Australian and Japanese patients, positive values for IgAIC were too infrequent to attempt a similar analysis, therefore we considered patients with IgAIC levels above the 50th percentile and those below. The two subgroups of Australian IgAGN patients, with IgAIC above the 50th percentile (12 patients) and below (nine patients) showed significantly different levels of IgA to dietary Ag (Table 1). Moreover, significant correlations were observed between IgAIC and IgA to dietary Ag values (OVA, r = 0.68, P
0.10

Italy

Austnlta

,cO.OOOl

p N5

~

pottents offected by IgAGN

c=J

control group

J

Japan • os

1 standard deyfetton

Fig 2. Circulating IgAIC in IgAGN patients and control groups (see legend to Fig 1).

In sera from Italian and Australian IgAGN patients, we observed that the values of IgA to various gluten fractions were significantly correlated (Table 2). IgA to albumins (OVA and BSA) were similarly correlated. Moreover, we found a significant correlation between the values of IgA to gluten fractions and IgA to heterologous albumins. These relationships were of lesser significance in Japanese sera (Table 2). Statistical analysis by means of chi-square test gave superimposable significances and were not reported in the tables. Lectin-binding IgA Activity

Increased lectin-binding IgA actIvIty in comparison to healthy controls was found in 0 to 72 %

484

COPPO ET AL Table 1.

Relationship Between IgA to Dietary Ag and IgAle Levels

Italian Patients

Australian Patients

IgA to Alimentary Ag OVA

Positive IgAIC (27 cases)

Negative IgAIC (36 cases)

0.56 ± 0.31 (0.26-1.71)

0.39 ± 0.24 (0.16-1.17)

BSA

0.54 ± 0.40 (0.07-2.1)

Glia-BB

0.47 ± 0.33 (0.18-1.55)

Glia-eth

0.42 ± 0.32 (0.18-1.77)

Glyc-gli

0.73 ± 0.65 (0.25-3.81 )

Glut

0.61 ± 0.33 (0.28-1.77)

IgAIC

0.92 ± 0.47 (0.52-2.45)

IgAIC >50th Percentile (12 cases) 0.51 ± 0.22 (0.25-0.96)

P<0.0007

IgAIC <50th Percentile (9 cases) 0.29 ± 0.13 (0.16-0.57)

IgAIC >50th Percentile (11 cases) 0.44 ± 0.18 (0.24-0.81)

P<0.003

0.48 ± 0.44 (0.16-1.99)

P<0.03

0.37 ± 0.27 (0.15-1.34)

0.54 ± 0.20 (0.25-0.95)

0.35 ± 0.37 (0.14-1.61)

P<0.008

0.51 ± 0.18 (0.27-0.78)

P<0.003

0.41 ± 0.10 (0.24-0.54)

P<0.009

0.52 ± 0.18 (0.24-0.88)

0.53 ± 0.40 (0.15-2.15)

0.62 ± 0.19 (0.32-0.88)

0.38 ± 0.11 (0.26-0.54)

0.24 ± 0.08 (0.09-0.43)

0.35 ± 0.10 (0.24-0.57)

0.54 ± 0.12 (0.39-0.81)

P<0.03

P<0.04

NS

0.50 ± 0.08 (0.40-0.64)

P<0.03

P<0.00003

0.31 ± 0.09 (0.18-0.50) 0.43 ± 0.11 (0.29-0.68)

P<0.03

0.40 ± 0.10 (0.27-0.65)

0.81 ± 0.30 (0.51-1.36)

0.31 ± 0.11 (0.20-0.60)

P<0.01

0.20 ± 0.04 (0.15-0.26)

P<0.02

0.74 ± 0.55 (0.26-2.71 )

0.40 ± 0.14 (0.16-0.74)

P<0.04

0.19 ± 0.06 (0.13-0.32)

0.40 ± 0.21 (0.10-0.85)

IgAIC <50th Percentile (14 cases) 0.29 ± 0.13 (0.15-0.77)

P<0.005

0.31 ± 0.16 (0.18-0.70)

0.42 ± 0.26 (0.11-1.02)

P<0.02

NS

Japanese Patients

0.45 ± 0.09 (0.34-0.70) NS

0.28 ± 0.11 (0.11-0.46)

0.79 ± 0.39 (0.53-1.77)

P<0.0001

0.22 ± 0.07 (0.12-0.37)

P<0.0003

NOTE. Italian IgAGN patients were divided according to positive or negative IgAIC data (above and below the 90th percentile), while Australian and Japanese were divided in two groups above and below the 50th percentile.

By comparing data from the three patient groups, we observed that Japanese patients displayed a significantly greater lectin-binding activity of IgA than Australian (Pokeweed I-mGlc, P<0.04; glia-eth GlcNAc, P
of sera from Japanese patients, 0 to 56% of Australians, and 6 % to 26 % of Italians. Sera from Italian and Australian patients did not display any increase in mean values of lectin-binding IgA activity when compared with the corresponding healthy control groups (Fig 3). No correlation was found with IgAlC values. Conversely, sera from Japanese IgAGN patients showed a significantly increased lectin-binding activity ofIgA versus Pokeweed lectin (P<0.0005), without correlation with IgAlC values (Fig 3). IgAlC values were significantly correlated with the lectin-binding IgA activity versus glia-eth (r 0.63, P
Relationship Among IgA Antibodies to Various Dietary Ag BSA

OVA It

Au

Ja

It

Au

Glia-BB Ja

It

Au

Gluten

Glia-eth Ja

It

Au

Ja

It

Au

Ja

BSA 0.70t 0.92:j: 0.66' Glia-BB 0.81' 0.90:j: 0.73' 0.84:j: Glia-eth 0.84' 0.85:j: 0.53t 0.78' 0.81:j: 0.49t 0.87' 0.95:j: 0.48t Glutenin 0.78:j: 0.83:j: 0.76' 0.85:j: 0.45t 0.73' 0.87:j: Glyc-gli 0.77:j: 0.44t 0.80' 0.82:j: 0.71:j: 0.88' 0.89:j: 0.61' 0.68' 0.84:j: 0.44t 0.80:j: Only significant correlations were reported: 'P< 0.005, tP< 0.05, :j:P< 0.01. Abbreviations: It, Italian IgAGN patients; Au, Australian IgAGN patients; Ja, Japanese IgAGN patients.

IGA TO DIETARY AG AND LECTIN-BINDING IGA

:::~.~~

.: c5

c§ & C§ c!J c!J

f::j ~ mm~ mm

...... .-.4 I.......

s..IMM 1-.,0"

IgAG" peltents

... IN_ 0 .. 1

lei..... l..o~

~ control I

att_tt. 1--0"

O"..-.tt. 0 .....

J I ,tend.rd dev.eUon

Fig 3. Lectin-binding IgA activities in IgAGN patients and control groups (see legend to Fig 1). The calculation of the ratio of lectin-binding IgA is described in Materials. P values are reported in Results.

mGlc v soybean gal, r = 0.52, P<0.05; soybean gal v glia-eth I-mGlc, r = 0.52, P<0.05; soybean gal v soybean I-mGlc, r = 0.53, P<0.05). Conversely, various lectin-binding activities were significantly correlated with levels of IgA antibodies to dietary Ag (Table 3). DISCUSSION

We report the analysis of sera from Italy, Australia, and Japan, showing different patterns of IgAIC and IgA against alimentary Ag and lectinbinding IgAs. Increased values of IgAIC were detected in about half of the Italian patients and in one quarter of Australians. Japanese IgAGN patients had an IgAIC prevalence the same as their controls. Mean values of IgAIC were only significantly increased in comparison to healthy controls in Italian patients. Positive values of IgA antibodies against dietary Ag had variable prevalences, particularly high in Italian and Australian patients. Mean values of

485

these antibodies were not significantly increased in any patient groups in comparison to the corresponding healthy populations. These data might be considered to demonstrate the lack of a positive role for alimentary Ag in IgA nephropathy, as suggested by several investigators who reported low levels of serum IgA to dietary in IgAGN patients, similar to those found in healthy controls. 9 ,l0,1315 However, a deeper analysis of the results we obtained gives some evidence that a subgroup of IgAGN patients presents a modified reactivity against alimentary Ag. Relationships exist between levels of IgAIC and IgA antibodies to dietary Ag, since patients with high levels of IgA to alimentary components also statistically present the highest values of IgAIC. This correlation is particularly evident in Italian and Australian patients. Moreover, the patients with positive data tend to have a cluster of increased levels of IgA against several alimentary Ag at the same time. A linear correlation is evident between various values of IgA to gluten fractions and heterologous albumins. Conversely, there were no such correlations in any of the three healthy control populations. Even though the relationship between serum IgAIC and IgA to dietary Ag does not prove their participation in the formation of mesangial immune deposits, the above findings are consistent with the hypothesis that in subgroups of IgAGN patients, a modified intestinal barrier permeability to alimentary components works in synergism with a mucosal immune system hyperreactivity, leading to an abnormal response to common foods. Several dietary elements, such as lectins 28 -3o and alcohoP7 can theoretically modify the intestinal permeability. Besides the role of antibodies, some IgA molecules in sera from the patients investigated were able to bind severallectins by a nonimmune bond, which could be inhibited by competitive sugars. Comparing patients' data with the corresponding controls, we found increased lectin-binding IgA activity in up to three quarters of Japanese patients, in half of Australians, and in one third of Italians. Only in Japanese patients' sera were mean values of lectin-binding IgA activity significantly increased in comparison to the healthy controls. Lectin-binding activity was not correlated with conglutinin-binding IgAIC levels, while a relationship was observed with IgA antibodies to alimen-

coppa

486 Table 3.

Relationship Between IgA Antibodies to Various Dietary Ag and Lectin-Binding IgA Activities Glia-eth 1-mGlc

Pokeweed 1-mGLc It

OVA BSA Glia-BB Glia-eth Glutenin Glyc-gli

ET AL

Au

Ja

It

Au

Glia-eth GlcNac Ja

0.39t 0.66' 0.62t O.S7t

0.42t

It

Au 0.62:j: 0.67' 0.S9t 0.6S:j: 0.70' 0.64t

Soya-eth 1-mGlc Ja

It

Au

Soybean Gal Ja

0.42t 0.42t 0.42t

It

Au

Soybean 1-mGlc Ja

It

Au

Ja

0.S2t 0.84' 0.43t 0.70' 0.76' 0.74' 0.S1t 0.S2t

Only significant correlations were reported: 'P< O.OOS, tP< O.OS, :j:P< 0.01.

tary Ag, suggesting a hypothetically common factor increasing the mucosal absorption of both Ag and lectins. The lectin-binding IgA activity we observed in some sera may be a consequence of increased exposure to environmental lectins of alimentary or infectious origin. Another possible explanation is that increased levels of naturally occurring or aberrantly glycosylated IgA molecules exist,38 bearing more or different sugar residues, thus increasing the binding to circulating lectins. Many IgA1 molecules from patients affected by IgAGN exhibit alterations in their oligosaccharide chain as compared with healthy controls (Davin JC, personal communication, 1989). The latter hypothesis presupposes a genetic influence on the production ofIgA with peculiar characteristics and/or a defect during the glycosi1ation process of IgA molecules in B lymphocytes. Following IgA reactions with lectins, IgA nonimmune complexes can form in circulation. Some macromolecular IgA have been found to be dissociated by sugar solutions, indicating a possible bond between IgA and lectins. 39 In conclusion, one may extrapolate from our data that in a subgroup of IgAGN patients, IgA antibodies to alimentary Ag may be involved in the formation of IgAIC. The prevalence and correlations of these serological abnormalities were particularly evident in Italian and Australian patients, more so than in the Japanese. This last group of patients was characterized by a high IgA capability to bind some lectins by a nonimmune mechanism. These data suggest the hypothesis that macromolecular IgA may circulate in IgAGN patients either in the form of immune complexes, at least in part as a response to alimentary Ag, or as nonimmune complexes due to IgA-Iectin interactions. The presence of lectin-binding sites in various glomerular structures 23 and cultured mesangial cells,24 and the capability of polymeric IgA of

binding to cultured mesangial cells, 24 may enhance the formation of glomerular deposits containing IgA. These factors could favor the renal delivery of circulating lectin-containing IgA macromolecules or IgAIC, as well as the in situ formation of immune and nonimmune glomerular IgA deposits. This is, of course, a hypothesis to be proved, since even though soybean Ag have recently been detected in mesangial deposits of IgAGN patients,9 particularly from Japan,40 no definite demonstration of diffuse deposition of alimentary Ag in renal tissue has been obtained. 41 Therefore, the relationship between the abnormalities in serum IgA and mesangial deposits in IgA nephropathy needs further investigation. ACKNOWLEDGMENT We would like to thank Dr Simon Koblar for help in revising the manuscript.

REFERENCES 1. Emancipator SN: Experimental models of IgA nephropathy. Am J Kidney Dis 5:415-419, 1988 2. Rifai A, Chen A, Imai H: Complement activation in experimental IgA nephropathy: An antigen-mediated process. Kidney Int 32:838-844, 1987 3. Coppo R: The pathogenetic potential of environmental antigens in IgA nephropathy. Am J Kidney Dis 5:420-424, 1988 4. Clarkson AR, Woodroffe AJ, Aarons I, et al: IgA nephropathy. Ann Rev Med 38:157-168, 1987 5. Tomino Y, Sakai H, Miura M, et al: Detection of antigenic substances in patients with IgA nephropathy. Contrib Nephrol 40:69-73, 1984 6. Jessen RH, Nedrud JG, Emancipator SN: A mouse model of IgA nephropathy induced by Sendai virus. Adv Exp Med Bioi 216:1609-1618, 1987 7. Emancipator SN, Gallo GR, Lamm ME: Experimental IgA nephropathy induced by oral immunization. J Exp Med 157:572-582, 1983 8. Genin G, Laurent B, Sabatier JC, et al: IgA mesangial deposits in C3H/HeJ mice after oral immunization with ferritin or bovine serum albumin. Clin Exp ImmunoI63:385-394, 1986

IGA TO DIETARY AG AND LECTIN-BINDING IGA 9. Russell M\V, Mestecky J, Julian BA, et al: IgA-associated renal disease: Antibodies to environmental antigens in sera and deposition of immunoglobulins and antigens in glomeruli. J Clin Immunol 6:74-86, 1986 10. Woodroffe AJ, Gormly AA, McKenzie PE, et al: Immunologic studies in IgA nephropathy. Kidney Int 18:366-374, 1980 11. Nagy J, Scott H, Brandtzaeg P: Antibodies to dietary antigens in IgA nephropathy. Clin Nephrol 29:275-279, 1988 12. Laurent J, Branellec A, Heslan JM, et al: An increase in circulating IgA antibodies to gliadin in IgA mesangial glomerulonephritis. Am J Nephrol 7:178-183, 1987 13. Fornasieri A, Sinico RA, Maldifassi P, et al: IgA-antigliadin antibodies in IgA mesangial nephropathy (Berger's disease). Br Med J 295:78-80, 1987 14. Rodriguez-Soriano J, Arrieta A, Vallo A, et al: IgA antigliadin antibodies in children with IgA mesangial glomerulonephritis. Lancet 1: 1109-1110, 1988 15. Coppo R, Roccatello D, Amore A, et al: Effects of a gluten-free diet in primary IgA nephropathy. Clin Nephrol 33:72-86, 1990 16. Sancho J, Egido J, Rivera F, et al: Immune complexes in IgA nephropathy: Presence of antibodies against diet antigens and delayed clearance of specific polymeric IgA immune complexes. Clin Exp Immunol 54: 194-202, 1983 17. Fornasieri A, Sinico RA, Maldifassi P, et al: Food antigens, IgA immune complexes and IgA mesangial nephropathy. Nephrol Dial Transplant 3:738-743, 1988 18. Sato M, Kojima K, Takayama K, et al: Glomerular deposition of food antigens in IgA nephropathy. Clin Exp Immunol 73:295-299, 1988 19. Roque-Barreira MC, Campos-Neto A: Jacalin: An IgAbinding lectin. J Immunol 134: 1740-1743, 1985 20. Damjanov I: Lectin cytochemistry and histochemistry. Lab Invest 57:5-20, 1987 21. Kondoh H, Kabayashi K, Hagiwara K, et al: Jacalin, a jackfruit lectin, precipitates IgAI but not IgA2 subclass on gel diffusion reaction. J Immunol Methods 80:171-173, 1986 22. Stockert RJ, Kressner MS, Collins JC: IgA interaction with the asialoglycoprotein receptor. Proc Nat! Acad Sci USA 79:6229-6235, 1982 23. Holthofer H: Lectin bindjng sites in kidney. A comparative study of 14 animals species. J Histochem Cytochem 31:531-537,1983 24. Amore A, Emancipator SN, Coppo R, et al: Specific binding of gliadin to rat mesangial cells in culture. Nephrol Dial Transplant 16:54, 1988 25. Roccatello D, Amprimo MC, Coppo R, et al: Influence of gluten derived fractions on chemiluminescence production by human neutrophils. J Biolumin Chemilumin 5:161-164, 1990 26. Davin JC, Derchenne C, Lombet J, et al: Acute experimental glomerulonephritis induced by the glomerular deposi-

487 tion of circulating polymeric IgA-concanavalin A complexes. Virchows Arch [AJ 415:7-20, 1989 27. Tomino Y, Sakai H, Miura M, et al: Effect of immunoglobulin depositions of glomerular sialic acids in patients with IgAGN nephropathy. Am J Nephrol 6:187-192, 1986 28. Douglas AP: The binding of a glycopeptide component of wheat gluten to intestinal mucosa of normal and coeliac human subjects. Clin Chern Acta 73:357-361, 1976 29. Nakata S, Kimura T: Effect of ingested toxic bean lectins on the gastrointestinal tract in the rat. J Nutr 115: 16211629, 1985 30. Lorenzsonn V, Olsen WA: In vivo responses of rat intestinal epithelium to intraluminal dietary lectins. Gastroenterology 82:838-848, 1982 31. Coppo R, Mazzucco G, Martina G, et al: Gluten induced experimental IgA glomerulopathy. Lab Invest 60:499506, 1989 32. D'Amico G: The commonest glomerulonephritis in the world: IgA nephropathy. Q J Med 245:709-727, 1987 33. Julian B, Waldo B, Rifai A, et al: IgA nephropathy, the most common glomerulonephritis worldwide. Am J Med 84: 129-132, 1988 34. Power DA, Simpson JG, Nicholls AJ, et al: IgA nephropathy is not a rare disease in the United Kingdom. Nephron 40: 180-184, 1985 35. Coppo R, Basolo B, Martina G, et al: Circulating immune complexes containing IgA, IgG and IgM in patients with primary IgA nephropathy and with Henoch-Schoenlein nephritis: Correlation with clinical and histologic signs of activity. Clin Nephrol 18;230-239, 1982 36. Coppo R, Basolo B, Roccatello D, et al: Immunological monitoring of plasma exchange in primary IgA nephropathy. Artif Organs 9:351-360, 1985 37. Worthington BS, Meserale L, Syrotuck JA: Effect of daily ethanol ingestion on intestinal permeability to macromolecules. Dig Dis 23:23-32, 1978 38. Monteiro RC, Chevailler A, Noel L-H, et al: Serum IgA preferentially binds to cationic polypeptides in IgA nephropathy. Clin Exp Immunol 73:300-306, 1988 39. Davin JC, Malaise M, Foidart J, et al: Anti-alpha-galactosyl antibodies and immune complexes in children with Henoch-Schoenlein purpura or IgA nephropathy. Kidney Int 31:1132-1139,1987 40. Sato M, Kojima H, Kino K, et al: Effects of neutral pepsin on the deposition of dietary antigens in glomeruli from IgA nephropathy. Clin Exp Immunol 81:137-141, 1990 41. Coppo R, Amore A, Roccatello D, et al: IgA to food and viral antigens in circulating immune complexes and mesangial deposits in IgA nephropathy (IgAGN). Abstracts of the XIth International Congress of Nephrology, Tokyo, Japan, July 1520, 1990, P 44A