Immunohistochemical and serological evidence for the role ofstreptococcal proteinase in acute post-streptococcalglomerulonephritis

GIL A. CU, SERGIO MEZZANO, JASON D. BANNAN, and JOHN B. ZABRISKIE

Laboratory of Clinical Microbiology and Immunology, Rockefeller University, and Division of Nephrology, Beth Israel Medical Center,New York City, USA; and Department of Nephrology, Faculty of Medicine, Universidad Austral de Chile, Valdivia, Chile

Immunohistochemical and serological evidence for the role ofstreptococcal proteinase in acute post-streptococcal glomerulone-phritis.

Background. We have previously demonstrated the preferentialsecretion of streptococcal proteinase or streptococcal pyrogenicexotoxin B (SPEB) by nephritic strains of Group A streptococciisolated from the skin or throat of patients with acute post-streptococcal glomerulonephritis (APSGN).

Methods. To further explore the possible role of SPEB inAPSGN, we performed ELISA studies to detect anti-SPEBantibodies in the sera of patients with APSGN, acute rheumaticfever (ARF), scarlet fever (SF) and normal children. UsingELISA, anti-SPEB titers on acute and convalescent APSGN serawere measured to determine immunity to APSGN. We alsoperformed immunofluorescence studies on APSGN and non-APSGN kidney biopsies to probe for the presence and localiza-tion of SPEB.

Results. Our data show that anti-SPEB antibodies are present inAPSGN sera and antibody titers are significantly higher than inARF, SF and normal sera. Anti-SPEB titers tend to rise acutelyand decrease with time but do not reach baseline after one year.When kidney biopsies were probed with rabbit anti-SPEB anti-body, 12 of 18 (67%) of the APSGN cases were positive while only4 of 25 (16%) of the non-APSGN cases were positive.

Conclusions. In summary, we were able to demonstrate uniquereactivity to SPEB in human sera and kidney biopsies of APSGNsuggesting a significant role of this toxin in the pathogenesis ofacute post-streptococcal glomerulonephritis.

Acute post-streptococcal glomerulonephritis (APSGN)has been considered to be an immune-complex mediatedkidney disease that has been strongly associated with throator skin infections with certain nephritogenic strains ofb-hemolytic Group A streptococci. Yet the exact mecha-

nism of how these nephritogenic strains specifically causethe disease remains to be elucidated. The acute onset of thedisease coupled with the fact that throat or skin infectionproduces the same kidney disease strongly suggests thatnephritogenic streptococci might secrete a product uniqueto those strains associated with nephritis.

Previous work in our laboratory has identified an extra-cellular plasmin-binding protein preferentially secreted bynephritis-associated Group A streptococci [1]. Amino-terminal sequencing of this 46 kD protein revealed that itwas structurally identical to streptococcal pyrogenic exo-toxin B (SPEB) and to a molecule previously described asstreptococcal proteinase, which is a cysteine protease foundin broth culture of Group A streptococci.

Streptococcal proteinase was first described by Elliot in1945 when he noted that Group A streptococci elaboratedan extracellular zymogen (molecular wt 42,000) in culturethat could be transformed into an active proteinase (mo-lecular wt 28,000) by proteolysis followed by reduction [2].The active enzyme has the ability to digest streptococcal Mprotein, casein, gelatin, human and rabbit fibrin. Underappropriate culture conditions, the zymogen may constituteover 95% of the extracellular proteins produced by GroupA streptococci. Reducing conditions (use of thiol com-pounds such as thioethanol, thioglycolic acids, cysteine or2,3-dimercaptopropanol) activates the zymogen by reduc-ing the disulfide bond that links the single half-cystineresidue in the protein to a methanethiol, a volatile mercap-tan, thus behaving like a sulfhydryl enzyme such as papain.The zymogen (42 kD) can be autocatalytically cleaved bythe reduced zymogen or by trypsin to become an activeproteinase. This results in the removal of about 100 aminoacids from the amino terminal end of the zymogen yieldinga protein of molecular wt 28 kD in size. In this regard,streptococcal proteinase behaves like a trypsin enzyme.Enzyme kinetic studies have shown that streptococcalproteinase has a high ratio of esterase to peptidase activity[3]. It is a cationic protein with a pk of 8.5.

Key words: Streptococcal pyrogenic exotoxin B, toxin, immune-complexmediated disease, infection, nephritis, plasmin, nephritogenic protein.

Received for publication September 3, 1996and in revised form March 13, 1998Accepted for publication April 16, 1998

© 1998 by the International Society of Nephrology

Kidney International, Vol. 54 (1998), pp. 819–826

819

Renewed interest in SPEB as an important toxin inhuman disease has demonstrated additional propertiessuch as its ability to induce high fever and rash, enhancehost susceptibility to lethal effects of other agents, andstimulate nonspecific T cell proliferation [4]. The activeenzyme can cleave the IL-1b precursor to the active IL-1 (amajor cytokine mediating inflammation), cleaves humanfibronectin and can degrade vitronectin (component of theextracellular matrix) [5, 6]. Furthermore, it cleaves mono-cytic cell urokinase receptor to release an active fragmentfrom the cell surface [7]. Passive and active immunizationof mice against SPEB appears to enhance survival againstvirulent strains of Group A streptococci [8]. These findingsare strongly suggestive that SPEB may be involved in thevirulence of Group A streptococci in the human host.

To further explore the possible role of SPEB in thepathogenesis of APSGN, we performed ELISA studies todetect the presence of anti-SPEB antibodies in the sera ofpatients with APSGN, acute rheumatic fever (ARF), scar-let fever (SF) and normal children in endemic areas. UsingELISA techniques, anti-SPEB titers on acute and conva-lescent sera of APSGN patients were measured to deter-mine possible immunity to APSGN. We also performedimmunofluorescence studies on kidney biopsy specimenson APSGN and non-APSGN patients to demonstrate itspresence and localization on human kidneys. The resultsclearly demonstrate that antibodies to SPEB are present inhuman sera of APSGN patients and may provide immunityto APSGN and that SPEB localizes in the kidneys whereactivation of complement and the inflammatory cascadesmay occur.

METHODSPurification of SPEB

Purified SPEB (in zymogen form) was obtained from theextracellular products of a nephritic strain of Group Astreptococcus F203D as previously described [9]. The bac-teria was initially grown in Todd-Hewitt and adapted tochemically-defined medium. After centrifugation to re-move the cells, the supernatant was passed through a45-micron filter and the filtrate was collected and precipi-tated with ammonium sulfate to 60% saturation. Theprecipitated protein was suspended in 50 mM Tris pH 8.2and dialyzed twice against 10 liters of similar buffer over-night to remove the ammonium sulfate. The dialyzedextracellular products were then passed through a DEAEcolumn and eluted with 0.2 M NaCl. Further purificationwas done using FPLC Mono Q column (Pharmacia BiotechInc., Pistacaway, NJ, USA). The purification procedure wasmonitored with Western blot using a rabbit polyclonalantibody against SPEB.

Human seraHuman sera were obtained from The Rockefeller Uni-

versity collection. Sera from children (ages 6 to 14 years of

age) with well-documented sequelae of Group A strepto-coccal infection like acute rheumatic fever (ARF), acutescarlet fever (SF) and acute post-streptococcal glomerulo-nephritis (APSGN) were used. Control sera were age-matched and were obtained from the same geographicalarea as the APSGN patients (Trinidad and Chile). Table 1summarizes the clinical and laboratory data on thesepatients. Children with ARF fulfilled the Jones Criteria forDiagnosis. APSGN patients had decreased complement,isolation of group A streptococci from skin lesions, ele-vated BUN, edema and raised ASO titers. SF patients hadclassical rash, pharyngitis, and in most cases isolation ofgroup A streptococci from the throat. Normal controls hadno signs of a streptococcal infection, negative throat cul-tures and low ASO titers. All sera were drawn during thefirst week of admission to the hospital or in the out-patientclinic, with the exception of serial bleeds on the thirteenAPSGN patients, who were followed for one year.

Animal sera

Rabbit polyclonal antibodies to purified SPEB (zymogenform) were produced using standard techniques [10].Briefly, 100 mg of purified SPEB was suspended in phos-phate buffered saline (PBS; pH 7.5) and mixed with equalvolume of complete Freund’s adjuvant (Difco Laborato-ries, Detroit, MI, USA) until a thick emulsion was ob-tained. The emulsion was injected subcutaneously into NewZealand White rabbits (5 to 7 lbs in weight) in five to sixdifferent sites. Two rabbits were used for immunization.Subsequent booster immunizations were done every twoweeks using 100 mg of purified SPEB mixed with incom-plete Freund’s adjuvant. Test bleeds were obtained everytwo weeks and antibody titers against SPEB were deter-mined using ELISA techniques (see below).

Kidney biopsies

Human kidney biopsies (APSGN and non-APSGNcases) were obtained from the Rockefeller University col-lection and Universidad Austral de Chile. All biopsy ma-terials were previously used to confirm the diagnosis ofAPSGN or other renal disease. APSGN cases (10 from skininfections and 8 from throat infections) were from patientsaged 8 to 23 while non-APSGN cases were from ages 14 to60. Due to paucity of control biopsies from the samegeographical area as the APSGN cases, we were compelled

Table 1. Clinical and laboratory data on patients with Group Astreptococcal infections, their sequelae and normal children

Category NumberAgeyears C3 values

AverageASO titers

APSGN 30 5–14 ,50 596ARF 20 5–18 NL 566SF 31 5–18 NL 448Controls 22 5–18 NL 160

Cu et al: Streptococcal proteinase in APSGN820

to use specimens obtained from different geographicalareas that are part of the University collection. All biopsysections were stored frozen at 270°C until use.

Immunofluorescence studies

Using a cryostat, frozen kidney biopsy tissues were cutinto 5 m slices and fixed in cold acetone for five minutes, airdried and washed in PBS (pH 7.4) for five minutes.Blocking with a protein blocking agent (Diftags Kit; Shan-don Inc., Pittsburgh, PA, USA) was done for 20 minutes.The sections were then incubated with prediluted anti-human C3-FITC (Shandon Inc.), prediluted anti-humanIgG-FITC (Shandon Inc.) or anti-SPEB antibodies (1:20and 1:40 dilution) at room temperature for 30 minutesinside a humidified chamber. For human complement andhuman IgG staining, direct immunofluorescent techniqueswere used. For the anti-SPEB staining, indirect immuno-fluorescence was done using a goat anti-rabbit IgG-FITCantibody (Sigma Chemical Company, St. Louis, MO, USA)as the secondary antibody (1:100 dilution) with incubationat room temperature for 30 minutes. The slides werewashed in PBS for five minutes three times and mountedbefore viewing with a fluorescence microscope. Preab-sorbed sera was prepared by incubating 1 mg of purifiedSPEB with equal volume of anti-SPEB sera at 37°C forthree hours. The insoluble material was removed by cen-trifugation at 12,000 rpm 3 15 minutes. The supernatantwas used at 1:20 dilution.

ELISA studies

Into a 96-well microtiter plate (Linbro/Titertek; ICNBiomedicals Inc., Horsham, PA, USA) purified SPEB (2mg/ml) in sample buffer (100 mM Tris, pH 9.8) was added at100 ml per well. The antigen was allowed to bind to the wellby incubating the sealed plate overnight at room tempera-ture. The plate was then washed for 10 minutes three timeswith ELISA wash buffer (10 mM Tris, 500 mM NaCl, 0.05%Brij 35). The primary antibody (anti-SPEB sera or humansera in varying dilutions of ELISA wash buffer) was addedand the plate incubated at 37°C for one hour underhumidified conditions. The plate was washed as before. Thesecondary antibody, goat anti-rabbit IgG-alkaline phospha-tase conjugate (Sigma Chemical Co.) or goat anti-humanIgG-alkaline phosphatase conjugate (Sigma Chemical Co.)diluted at 1:1000 in ELISA wash buffer, was then addedand the plate incubated at 37°C for one hour underhumidified conditions. After washing the plate as before,100 ml of the phosphatase substrate solution [Sigma 104phosphatase tablets dissolved in substrate buffer (1 M Tris,pH 9.8, 1 mM MgCl2) at 1 mg/ml] was added to each welland incubated at 37°C for one hour. The plate was thenread at 405 nm absorbance using a Dynatech MR 4000ELISA reader (Dynatech Laboratories, Chantilly, VA,USA). Absorbance readings were corrected against a sam-ple buffer blank, antigen control and antibody control.

Human sera were first diluted 1:100 and then furtherdiluted depending on whether the O.D. reading was .1.0.

Complement activation

Total hemolytic complement in the serum was measuredby a standard assay kit for total complement. Venousspecimens of clotted blood (human and rabbit) were col-lected and placed on ice at 4°C for one hour, centrifuged at4°C at 2,000 RPM for 10 minutes, the serum removed andstored at 270°C until use. On the day of the experiment,the sera were thawed and the control serum (200 ml) wasmixed 1:1 with 200 ml saline. The experimental sample (200ml) was mixed 1:1 with 200 ml of SPEB containing 200 mg ofSPEB. The samples were incubated for one hour at 37°C,centrifuged at 5,000 RPM at 4°C for 10 minutes andassayed for total complement levels.

Statistical analyses

For the ELISA studies comparing the level of anti-SPEBtiters among different groups of Group A streptococcalinfections (AGN, ARF, SF and normal children), theone-way ANOVA test was used with a 5 0.05 [11]. The nullhypothesis assumes that all the groups have similar levels ofanti-SPEB titers. For comparing the level of anti-SPEBtiters between the convalescent sera of APSGN patients(after 1 year) and normal children, the t-test was used witha 5 0.05. For the biopsy studies, the chi-square test [11]was used to analyze the outcome of the immunofluorescentstaining using the anti-SPEB antibody. The null hypothesisassumes that all groups have the same frequency of positiveand negative staining. The level of significance was set at a5 0.05. All the statistical calculations were done using SPSSStatistical Package (SPSS Inc., Chicago, IL, USA).

RESULTS

Comparative anti-SPEB titers between APSGN, ARF, SFand normal sera

Human sera from acute cases of acute post-streptococcalglomerulonephritis (APSGN, N 5 30), acute rheumaticfever (ARF, N 5 20), scarlet fever (SF, N 5 31) and normalchildren (N 5 22) were tested for the presence and level ofanti-SPEB antibodies using ELISA techniques. Table 1summarizes the clinical and laboratory values for eachgroup of patients. Results show that significant anti-SPEBtiters were found in the sera of APSGN (mean O.D. 5 700)patients compared to the other groups (ARF, mean O.D. 50.100; SF, mean O.D. 5 0.025, and normal children, meanO.D. 5 0.025). P value was ,0.001 (Fig. 1A).

Figure 1B summarizes the anti-streptolysin O (ASO)titers in each patient group in comparison to the anti-SPEBtiters in these patients. As can be seen, all streptococcalinfection groups had significantly elevated ASO titers com-pared to controls but were not significantly different fromeach other.

Cu et al: Streptococcal proteinase in APSGN 821

When the sera of 13 documented cases of APSGN werefollowed for 52 weeks, serial anti-SPEB titers tended to riseduring the first week and peaked during the first andsecond weeks following diagnosis. Thereafter, the titersdecreased but remained significantly higher than normalcontrols at the end of 52 weeks (Figs. 2 and 3). The P valuewas , 0.0001.

Immunofluorescence studies on kidney biopsies

Eighteen APSGN and 25 non-APSGN kidney biopsyspecimens were examined for the presence of SPEB usingan indirect immunofluorescence technique. All the APSGNcases were from school-aged children, but due to paucity ofnon-APSGN cases in children, adult biopsy cases wereobtained as non-APSGN control group. The latter casesinclude four normal kidney tissue, 3 membranous nephrop-athy, 2 IgA nephropathy, 1 crescentic glomerulonephritis, 3minimal change nephropathy, 1 HIV nephropathy, 1 inter-

stitial nephritis, 3 lupus nephritis, 1 ischemic renal diseasedue to renal artery stenosis, 3 hypertensive nephrosclerosis,1 thin basement membrane disease and 2 diffuse glomer-ulosclerosis.

All the APSGN kidney biopsies showed extensive depo-sition of C3 in the mesangial and glomerular capillaryareas, and weak to minimal deposition of human IgG andIgM. Twelve of the 18 APSGN cases (67%) had positiveimmunofluorescence for SPEB when the rabbit polyclonalanti-SPEB sera was used as the primary antibody. Figure 4shows the staining pattern of human C3, IgG, IgM, SPEB,rabbit sera preabsorbed with purified SPEB and staining ofa non-APSGN kidney biopsy section. The staining patternwith the anti-SPEB shows coarsely granular staining in themesangial and glomerular capillary areas. Four of the 25non-APSGN biopsies (16%) were positive for SPEB. Inter-estingly, three of the four non-APSGN cases were mem-branous nephropathy and one case with IgA nephropathy.None of the four non-APSGN cases that stained with theanti-SPEB had clinical or serological evidence of recentstreptococcal infection. Preabsorption of the anti-SPEBsera with purified SPEB abolished the staining for SPEB.No staining of any of the biopsies was seen with preimmunerabbit sera. Table 2 summarizes the results. Using thechi-square analysis to determine the outcome of the stain-ing, a significant difference was noted among the groups(P 5 0.0007).

Complement activation studies

Suspecting that SPEB might activate the complementpathway in the absence of antibody, human and rabbit seraknown to contain little or no antibodies to SPEB weremixed with purified SPEB (1 mg/ml) in a 1:1 ratio asdescribed in the Methods section. Table 3 summarizes theresults, and it is clear that SPEB activates complement inthe absence of detectable antibody. These results stronglysuggest that SPEB activates complement via the alternativepathway.

DISCUSSION

In this report, we demonstrated the presence of signifi-cant anti-SPEB titers in APSGN patients compared toother inflammatory sequelae of Group A streptococcalinfection (acute endemic areas of APSGN). We have alsoshown in a group of 13 documented cases of APSGN thatthe anti-SPEB titers tend to rise and peak during the firsttwo weeks of onset of disease and then lower with time.However, the titers do not reach baseline levels of controlsand remained significantly higher even after a year offollow-up (Figs. 2 and 3). These findings correlate quitewell with the well known clinical observation that recur-rences of APSGN are extremely rare and one attack usuallyconfers life long immunity.

Using indirect immunofluorescence techniques, 67% ofAPSGN kidney biopsy specimens stained for SPEB while

Fig. 1. (A) Anti-SPEB titers in sera from patients with acute post-streptococcal glomerulonephritis (APSGN), acute rheumatic fever (ARF),scarlet fever (SF) and normal children. *P , 0.001. (B) ASO titers fromthe same group of patients, indicating a definite serological response to astreptococcal extracellular product in the streptococcal infections groups.*P , 0.001 for AGN, ARF and SF vs. normals. P 5 NS vs. AGN, ARF andSF.

Cu et al: Streptococcal proteinase in APSGN822

only 16% of non-APSGN biopsies were positive for SPEB.The diffuse granular staining of the mesangial and endo-capillary areas of the glomeruli suggests the possible affin-ity of the antigen to these areas of the glomeruli. Theseimmunofluorescent findings coupled with the serologicalreactivity to SPEB strongly suggest a causal role for thisprotein in the pathogenesis of APSGN.

When our results are compared to other reports concern-ing the immunological response to SPEB, certain discrep-ancies appear. While it is well known that all group Astreptococci contain the gene for SPEB [12], Poon-King etal clearly showed that nephritis associated strains prefer-entially secreted the SPEB protein in the extracellularproducts when compared to non-nephritis associatedstrains [1]. Thus, expression of the SPEB protein in thisparticular strain may be crucial to the development ofAPSGN and the subsequent immunological response tothis antigen in these patients. Support for this concept isfound in the work of Vogt et al, who also found a uniqueserological response to their cationic protein (now known

to be SPEB) in APSGN patients compared to normals andrheumatic fever patients [13].

In contrast to these results, both Holm et al [14] andNorrby-Telgrud et al [15] have noted the presence ofsignificant levels of SPEB antibodies in the sera of normalpatients, patients with tonsillitis, women in the first trimes-ter of pregnancy, and patients with streptococcal toxicshock syndrome. The fact that the majority of their patientswere adults (age range 21 to 55) and our patients primarilywere children (5 to 15 years old) might explain the fact thatantibodies to SPEB were more often seen in adult patientsirrespective of the infection. It might also explain the wellknown clinical observation that APSGN is primarily adisease of childhood, since normal children of this agegroup have generally low titers to this streptococcal pro-tein.

Acute post-streptococcal glomerulonephritis (APSGN)had been classically described as an immune-complex renaldisease following a Group A streptococcal throat or skininfection. Although certain serological types of Group Astreptococci have been strongly associated with this disease[16], not all strains within a given nephritogenic type causenephritis. This suggests that a unique streptococcal productis being elaborated only by certain strains (irrespective ofthe M protein serotypes) originating from the of skin orthroat, which may lead to the clinical and pathologicalpicture seen in APSGN.

Numerous investigations for the past decades have fo-cused on identifying and characterizing the nature of thestreptococcal antigen(s) that may be responsible for incit-ing the inflammation of the kidneys. Andres et al identifieda streptococcal antigen(s) in glomeruli of APSGN patientsthat appeared to be from the streptococcal cell wall [17].Lange et al identified a cytoplasmic antigen called en-dostreptosin in glomeruli [18], and antibodies to this anti-gen were significantly higher in sera from patients [19]. Animportant and often overlooked point is that all investiga-tors noted that biopsies taken early during the course of thedisease rarely revealed the presence of antibody but did

Fig. 2. Serial anti-SPEB titers in the sera ofpatients with APSGN. N 5 13 cases.

Fig. 3. Comparison of anti-SPEB titers between convalescent APSGNsera (after 1 year) and normal sera. P , 0.0001 by the t-test.

Cu et al: Streptococcal proteinase in APSGN 823

exhibit large deposits of complement and antigen alone [13,18].

Most important to the present studies were a series ofobservations by Vogt and colleagues [13] in which theyreported the presence of a streptococcal extracellular prod-uct they called “the cationic extracellular streptococcalantigen.” This antigen was localized in 8 of 18 kidneybiopsies from patients with APSGN. Significantly elevatedlevels of antibodies to this cationic antigen were alsopresent in the sera of APSGN patients. Vogt later acknowl-edged that this cationic antigen was structurally identical tostreptococcal proteinase or SPEB (personal communica-tion). Yoshizawa et al also reported a 43 kD pre-absorbingstreptococcal antigen (PA-Ag) identified in kidney biopsiesthat appeared to activate the alternative complement path-way [20]; However, a partial sequence of this antigen

revealed no homology with SpeB or streptococcal protein-ase.

In this laboratory, Villarreal et al identified a nephritisstrain-associated protein present in the extracellular prod-ucts of nephritic strains of Group A streptococci [21]. Usinga rabbit polyclonal antibody against this 46 kD protein, theydemonstrated positive staining in 14 of 21 APSGN kidneybiopsies and negative staining in 15 non-APSGN biopsies.Serological studies also revealed that APSGN patient’s serareacted preferentially with the antigen, while only 20% ofacute rheumatic fever and impetigo sera were positive [22].Subsequent studies by Johnston and Zabriskie [9] identifieda similar protein from the extracellular products of thesame nephritic strains used by Villarreal et al. Aminoterminal sequencing of the presumed nephritis strain-associated protein revealed it to be a streptokinase and

Fig. 4. Immunofluorescent staining of kidney biopsy section of one case of APSGN using different antibodies (340). (A) Extensive deposition of C3in mesangium and glomerular capillary walls. (B and C) Minimal staining for IgG and IgM. (D) shows extensive deposition of SPEB in mesangium andglomerular capillary wall. Staining for SPEB can be abolished by pre-absorbing the rabbit sera against purified SPEB (E). (F) The absence of SPEB ina non-APSGN kidney biopsy section. (case of IgA nephropathy).

Table 2. Immunofluorescent results of staining for SPEB in APSGNand non-APSGN biopsies

Biopsyspecimens Anti-SPEB (1) Anti-SPEB (2) Total

APSGN 12 (67%) 6 (33%) 18Non-APSGNa 4 (16%) 21 (84%) 25

P 5 0.0007 (chi-square analysis). Abbreviations are: SPEB, streptococ-cal proteinase; ASPGN, acute post-streptococcal glomerulonephritis.

a Defined as normal kidney tissue, membranous nephropathy, IgAnephropathy, vasculitis, minimal change, lupus nephritis, nephrosclerosis,ischemic renal disease, crescentic GN, interstitial nephritis, or thinbasement membrane disease

Table 3. Total hemolytic complement levels in human and rabbit serawith/without the addition of streptococcal pyrogenic exotoxin B (SPEB)

GroupSPEBtiter Mixture

Complementlevel % Change

I. Humana) 200 ml 20 Saline 200 ml 36 mg % 0%b) 200 ml 20 SPEBa 200 ml 18 mg % 250%

II. Rabbita) 200 ml 0 Saline 200 ml 20 mg % 0%b) 200 ml 0 SPEBa 200 ml , 10 mg %b .250%a Contained 200 mg SPEBb Less than dectable range

Cu et al: Streptococcal proteinase in APSGN824

biochemical analysis confirmed its ability to act as plasmin-ogen activator. Mezzano et al, however, failed to demon-strate unique reactivity to streptococcal streptokinase ineither the sera or kidney biopsies of patients with APSGN[23]. This suggested that Villarreal’s nephritis strain-asso-ciated protein was immunologically different from strep-tokinase.

Because of these discrepancies, Poon-King et al re-examined the extracellular product of the same nephriticstrains of Group A streptococci that Villarreal andJohnston studied and isolated a 46 kD protein band onSDS-PAGE gels unique to these strains [1]. Amino termi-nal sequencing of the purified protein revealed it to be100% homologous to the precursor of streptococcal pyro-genic exotoxin B (SPEB) and identical to Elliot’s previouslydescribed streptococcal proteinase [24, 25]. In the samestudy, the purified protein had the additional property ofbeing able to bind human plasma in vitro (and hence isalternatively called nephritis plasmin binding protein).

The fact that both Dr. Vogt and our group have identi-fied a cationic protein present in the glomeruli of APSGNpatients raises the intriguing concept of the mechanism ofactivation of this antigen in the pathway of the disease. Asoriginally proposed by Batsford et al [25], they suggestedthat their cationic protein (now known to be streptococcalproteinase) could penetrate the negatively charged base-ment membrane (irrespective of its size) and deposit intothe subepithelial area where it might behave as a “plantedantigen” with subsequent activation of the complementsystem either alone or later as an antigen-antibody com-plex. It has been known that cationized macromolecules(such as ferritin, 500 kD) can penetrate the negativelycharged basement membrane and it was determined thatantigens of 40 to 1000 3 103 daltons in size and with pIgreater than 8.5 to 9.0 can accumulate in the glomerularbasement membrane (GBM) [25]. On the other hand,soluble immune complexes with net overall positive chargeshave been demonstrated in the GBM while anionic orneutral immune complexes tend to remain in the capillaryside of GBM or in the mesangium [26–31].

Our preliminary observations of the activation of thealternate complement pathway by SPEB adds furtherstrength to the “planted antigen” hypothesis. The activa-tion of the complement cascade via the alternate pathwaywould explain the observation by both Vogt and our groupthat complement deposition and antigens appear in theglomeruli long before the appearance of antibody. Subse-quently, SPEB in combination with antibody could alsodeposit complexes in the kidney, or antibody alone couldbind to the foreign planted antigen. Elution studies ofglomerular sections should be able to determine whetherSPEB is still present in the glomeruli after elution ofantibody. The studies would be similar to those carried byDixon, Oldstone and Tionetti in experimental immunecomplex disease [32].

The other theory is related to the plasmin-binding prop-erties of SPEB. Group A streptococci have been known toproduce a streptokinase that activates plasminogen to theactive plasmin [33]. This active plasmin when bound toSPEB is not inactivated by a-antiplasmin. This couldprovide a mechanism for invasion of normal tissue barrierswhereby unregulated and excess plasmin in the circulationmay activate complement and inflammation [34]. Boththeories, albeit interesting, will need further experimenta-tion.

While the exact mechanism whereby streptococcal pyro-genic exotoxin B (SPEB) alone or in combination withplasmin initiates the pathology seen in APSGN, the factthat SPEB plays an important role in APSGN is strength-ened by the following observations. First, it is expressed inall nephritic strains noted to date [1]. Secondly, there is amajor serological reactivity to the protein in the sera ofAPSGN patients. Thirdly, its presence has been detected ina high percentage of APSGN biopsies. Finally, at least twoindependent investigators have identified the antigen inAPSGN biopsies as SPEB. The molecular weight of theprotein (;43 to 46 kD), its serological reactivity in APSGNsera and its presence in the biopsies strongly suggest thatthe original nephritis strain associated protein, Vogt’scationic protein and our nephritis plasmin binding proteinor SPEB (Table 4) are in reality the same nephritogenicprotein and plays a causal role in the pathogenesis of acutepost-streptococcal glomerulonephritis.

ACKNOWLEDGMENTS

This work was supported in part by National Institutes of Health grant1R01 DK-41275-01, General Clinical Research Center Grant MO1 andFondecyt 194/0895.

Reprint requests to Gil A. Cu, M.D., Laboratory of Clinical Microbiologyand Immunology, Rockefeller University, 1230 York Ave., New York, NewYork 10021, USA.

REFERENCES1. POON-KING R, BANNAN J, VITERI A, CU G, ZABRISKIE JB: Identifica-

tion of an extracellular plasmin binding protein from nephritogenicstreptococci. J Exper Med 178:759–763, 1993

2. ELLIOTT SD: A proteolytic enzyme produced by Group A streptococciwith special reference to its effect on the type-specific M antigen. JExp Med 81:573–592, 1945

3. LIU TY: Demonstration of the presence of a histidine residue at theactive site of streptococcal proteinase. J Biol Chem 242:4029–4032,1967

Table 4. Structural, serological and immunological comparisons ofnephritogenic proteins

NSAP(Villarreal,1979 [18])

Cationic Ag(Vogt, 1983

[16])

NPBP/SPEB(Poon-King,

1993 [1])

Molecular wt 46 kD 43 kD 43 kDpI Not done 8.6–8.8 Not doneAntibody in sera 1 1 1Biopsy 1 1 1

Cu et al: Streptococcal proteinase in APSGN 825

4. HAUSER AR, STEVENS DL, KAPLAN EL, SCHLIEVERT PM: Molecularanalysis of pyrogenic exotoxins from Streptococcus pyogenes isolatesassociated with toxic shock-like syndrome. J Clin Microbiol 29:1562–1567, 1991

5. KAPUR V, MAJESKY MW, LI LL, BLACK RA, MUSSER JM: Cleavage ofinterleukin 1 beta (IL-1 beta) precursor to produce active IL-1 beta bya conserved extracellular cysteine protease from Streptococcus pyo-genes. Proc Natnl Acad Sci U S A 90:7676–7680, 1993

6. KAPUR V, TOPOUZIS S, MAJESKY MW, LI LL, HAMRICK MR, HAMILLRJ, PATTI JM, MUSSER JM: A conserved Streptococcus pyogenesextracellular cysteine protease cleaves human fibronectin and de-grades vitronectin. Microbial Pathogen 15:327–346, 1993

7. WOLF BB, GIBSON CA, KAPUR V, HUSSAINI IM, MUSSER JM, GONIASSL: Proteolytically active streptococcal pyrogenic exotoxin B cleavesmonocytic cell urokinase receptor and releases an active fragment ofthe receptor from the cell surface. J Biol Chem 269:30682–30687, 1994

8. KAPUR V, MAFFEI JT, GREER RS, LI LL, ADAMS GJ, MUSSER JM:Vaccination with streptococcal extracellular cysteine protease (inter-leukin-1 beta convertase) protects mice against challenge with heter-ologous group A streptococci. Microbial Pathogen 16:443–450, 1994

9. JOHNSTON KH, ZABRISKIE JB: Purification and partial characteriza-tion of the nephritis strain-associated protein from Streptococcuspyogenes, group A. J Exper Med 163:697–712, 1986

10. Immunizations, in Antibodies: A Laboratory Manual, edited by HAR-LOW E, LANE D, Cold Spring Harbor, Cold Spring Harbor LaboratoryPress, 1988, pp 53–38

11. BAILAR J, MOSTELLER F: Medical Uses of Statistics (2nd ed). Boston,NEJM Books, 1992, p 449

12. YU CE, FERRETTI JJ: Frequency of the erythrogenic toxin B and Cgenes (SpeB and SpeC) among clinical isolates of group A strepto-cocci. Infec Immun 59:211–215, 1991

13. VOGT A, BATSFORD S, RODRIGUEZ-ITURBE B, GARCIA R: Cationicantigens in poststreptococcal glomerulonephritis. Clin Nephrol 20:271–279, 1983

14. HOLM SE, NORRBY A, BERGHOLM AM, NORGREN M: Aspects ofpathogenesis of serious group A streptococcal infections in Sweden1988–1989. J Infect Dis 166:31–37, 1992

15. NORRBY-TEGLUND A, PAUKSENS K, HOLM SE, NORGREN M: Thereation between low capacity of human sera to inhibit streptococcalmitogens and serious manifestations of disease. J Infect Dis 170:585–591, 1996

16. STOLLERMAN GH: Nephritogenic and rheumatogenic group A strep-tococci. J Infect Dis 120:258–263, 1969

17. ANDRES GA, ACCINNI L, HSU KC, ZABRISKIE JB, SEEGAL BC:Electron microscopic studies of human glomerulonephritis with fer-ritin-conjugated antibody. Localization of antigen-antibody complexesin glomerular structures of patients with acute glomerulonephritis. JExper Med 123:399–412, 1966

18. LANGE K, SELIGSON G, CRONIN W: Evidence for the in situ origin ofpoststreptococcal glomerulonephritis: Glomerular localization of en-dostreptosin and the clinical significance of the subsequent antibodyresponse. Clin Nephrol 19:3–10, 1983

19. SELIGSON G, LANGE K, MAJEED HA, DEOL H, CRONIN W, BOVIE R:Significance of endostreptosin antibody titers in poststreptococcalglomerulonephritis. Clin Nephrol 24:69–75, 1985

20. YOSHIZAWA N, OSHIMA S, SAGEL I, SHIMIZU J, TRESER G: Role of astreptococcal antigen in the pathogenesis of acute poststreptococcalglomerulonephritis. Characterization of the antigen and a proposedmechanism for the disease. J Immunol 148:3110–3116, 1992

21. VILLARREAL H JR, FISCHETTI VA, VAN DE RIJN I, ZABRISKIE JB: Theoccurrence of a protein in the extracellular products of streptococciisolated from patients with acute glomerulonephritis. J Exper Med149:459–472, 1979

22. OHKUNI H, FRIEDMAN J, VAN DE RIJN I, FISCHETTI VA, POON-KING T,ZABRISKIE JB: Immunological studies of post-streptococcal sequelae:Serological studies with an extracellular protein associated withnephritogenic streptococci. Clin Exper Immunol 54:185–193, 1983

23. MEZZANO S, BURGOS E, MAHABIR R, KEMENY E, ZABRISKIE JB:Failure to detect unique reactivity to streptococcal streptokinase ineither the sera or renal biopsy specimens of patients with acutepoststreptococcal glomerulonephritis. Clin Nephrol 38:305–310, 1992

24. LIU TY, ELLIOT SD: Streptococcal proteinase, in The Enzymes (3rded), New York, Academic Press, 1971

25. BATSFORD S, OITE T, TAKAMIYA H, VOGT A: Anionic binding sites inthe glomerular basement membrane: Possible role in the pathogenesisof immune complex glomerulonephritis. Renal Physiol 3:336–340,1980

26. GALLO GR, LAMM ME: Glomerular deposition diseases: Insights intomechanisms. Uremia Invest 8:189–192, 1984

27. GALLO GR, CAULIN-GLASER T, EMANCIPATOR SN, LAMM ME: Ne-phritogenicity and differential distribution of glomerular immunecomplexes related to immunogen charge. Lab Invest 48:353–362, 1983

28. GALLO GR, CAULIN-GLASER T, LAMM ME: Charge of circulatingimmune complexes as a factor in glomerular basement membranelocalization in mice. J Clin Invest 67:1305–1313, 1981

29. CAULIN-GLASER T, GALLO GR, LAMM ME: Nondissociating cationicimmune complexes can deposit in glomerular basement membrane. JExper Med 158:1561–1572, 1983

30. GALLO GR, CAULIN-GLASER T, LAMM ME: Role of electrostaticcharge interactions in glomerular deposition of immune complexes.Pathobiol Ann 12:203–211, 1982

31. CAULIN-GLASER T, EMANCIPATOR SN, GALLO GR, LAMM ME:Charge-related deposition of immune complexes in the glomerularbasement membrane is independent of Fc effector function. Am JPathol 119:288–293, 1985

32. DIXON FJ, OLDSTONE MBA, TIONETTI G: Pathogenesis of immunecomplex glomerulonephritis of New Zealand mice. J Exper Med134:655–713, 1971

33. LOTTENBERG R, DESJARDIN LE, WANG H, BOYLE MD: Streptokinase-producing streptococci grown in human plasma acquire unregulatedcell-associated plasmin activity. J Infect Dis 166:436–440, 1992

34. FUJIGAKI Y, BATSFORD SR, BITTER-SUERMANN D, VOGT A: Comple-ment system promotes transfer of immune complex across glomerularfiltration barrier. Lab Invest 72:25–33, 1995

Cu et al: Streptococcal proteinase in APSGN826