Synergistic Pathogenic Effects of Combined Mouse Monoclonal Anti-Desmoglein 3 IgG Antibodies on Pemphigus Vulgaris Blister Formation (1)
Conforma- tional epitope mapping in model mice revealed that anti- Dsg3 IgG were predominantly raised against the middle to C-terminal extracellular domains of mouse Dsg3, where amino-acid residues are less conserved among desmoglein isoforms. PV model mice receiving naive splenocytes produced antibodies against the N-terminal domain of Dsg3 more frequently than mice receiving immunized splenocytes.
All of the NAK mAbs had the k isotype light chain. Four (NAK1, 2, 3, 5) had IgG1 and six (NAK4, 7, 8, 9, 10, 11) had IgG2a heavy chains. With ELISA, all of the mAbs reacted only with mouse Dsg3 and not with mouse Dsg1. Of these mAbs, NAK1 and NAK9 cross reacted with human Dsg3 but not with human Dsg1, and recognized the N-terminal residues 1–162. these NAK mAbs recognize different epitopes on Dsg3, and six of them recognize the N-terminal 1–162 residues.
NAK1, 4, 7, 8, 9, 10 mAbs precipitated residues 1–162 and 1–402, but not residues 195–565 or 403–565 (Figure 2a and b). Their epitopes appear to reside in residues 1–162, the N-terminal portion of the extracellular domain of mouse Dsg3. NAK2 and NAK5 reacted with residues 1–402 and 195–565, but not with residues 1–162 or 403–565. NAK3 precipitated residues 195–565 and 403–565, but not residues 1–162 or 1–402. Its epitope appears to reside in residues 403–565, the C-terminal portion of the extracellular domain. NAK11 mAb precipitated only residues 1–402. NAK1 mAb maintained its reaction with human Dsg3 mutated at T31, K33, I34 (Dsg3-M2), V53-N56 (Dsg3-M3), E70, S73 (Dsg3-M4), L75, T77 (Dsg3-M5), A83, and Q84 (Dsg3-M6), but lost reactivity when the Dsg3 specific residues T25, Y28, or Q29 were mutated (Dsg3-M1, M1-2, M1-2-3). NAK1 gained recognition of human Dsg1 modified with Dsg3-specific residues at T25, Y28, or Q29 (Dsg1-M1-2, M1-2-3). Similarly, NAK9 mAb reacted with human Dsg3 mutated at Dsg3-specific residues T31, K33, I34 (Dsg3-M2), E70, S73 (Dsg3-M4), L75, T77 (Dsg3-M5), A83, and Q84 (Dsg3-M6), but lost recognition when either T25, Y28, Q29 (M1), or V53-N56 (Dsg3-M3) were mutated. NAK9 gained recognition of human Dsg1 when mouse residues T25, Y28, Q29, and V53-N56 were introduced (Dsg1-M1-2- 3), but not when T25, Y28, and Q29 were introduced with T31, K33, and I34 (Dsg1-M1-2).
When individual NAK mAbs were injected, none of them induced gross blistering; this was expected, as Dsg1 co-expression in the skin of neonatal mice compensates for the impaired adhesive function of Dsg3. However, all of the mAbs except NAK3 induced varying degrees of microscopic blistering, with histological evidence of suprabasilar acantholysis. Neonatal mice that were co-injected with each of the NAK mAbs (except NAK3 and NAK5) developed extensive gross blisters with suprabasilar acantholysis, while mice injected with the PF IgG alone did not show any blister formation.
With inoculating individual hybridoma cells intraperitoneally, none of the recipient Rag2-/- immunodeficient mice (6–8 weeks old) developed this phenotype, even after obvious ascites fluid formation at day 15. Although all the mice showed in vivo IgG deposition on keratinocyte cell surfaces of the skin and mucous membranes, no blister formation was apparent in the oral mucosa at the histological level. When hybridoma cells for NAK1, 2, 4, 7, 8, 9, 10, and 11 were combined, the recipient mice developed weight loss, patchy hair loss, and crusted erosions around the snout, at approximately days 10–15 after the inoculation. The development of the PV phenotype was observed even after mAbs NAK8, NAK4, NAK10, and NAK9 were sequentially removed from the combination. The minimum combination that was sufficient to induce the PV phenotype in all recipient mice tested was NAK1, NAK2, NAK7, and NAK11, or NAK2, NAK3, NAK5, and NAK11, each of which recognize different epitopes on mouse Dsg3. The minimum combination that could induce the PV phenotype of at least one of the mice tested was NAK2 and NAK11 hybridoma cells.
With an in vitro dissociation assay using primary cultured mouse keratinocytes, NAK3 and NAK5, which failed to induce apparent pathogenic activity in the passive transfer model, gave smaller dissociation indexes than the others (0.5 and 12.5%, respectively), while NAK9 had the highest dissocia- tion index (45.1%). When several NAK mAbs were combined, without changing the total amount of IgG added (1 ug/ml), dissocia- tion scores greatly increased. When NAK1, 2, 4, 7, 9, 10, and 11 mAbs were combined, the dissociation score was equivalent to that of the AK23 mAb.
In vitro dissociation assay
To evaluate the activity of mAbs in the inhibition of keratinocyte cell–cell adhesion, we modified a previously described in vitro dissociation assay. For these cultures, skin specimens prepared from neonatal ICR mice at 12–24 hours of age were incubated in dispase II, and separated epidermis samples were further incubated in 0.25% Trypsin. Isolated keratinocytes were dispensed into 12-well culture plates with keratinocyte culture medium CnT-02. When keratinocytes were confluent, 1.2mM calcium was added and cells were incubated for 24hours. Recombinant exfoliative toxin A (0.25 mg/ml) produced in E. coli, which specifically digests mouse Dsg1, was added to cultures 2 hours before the assay, in which 1 mg/ml of individual or pooled NAK mAbs (1 mg/ml total) were added to culture media. After washing with 0.9 mM Ca2 þ -phosphate-buffered saline twice, mouse keratinocytes were incubated with dispase II for 15 minutes to release cells as sheets. Released sheets were carefully washed twice with 0.9 mM Ca2 þ -phosphate-buffered saline and subjected to mechanical stress by pipetting with a 1ml disposable pipette tip. Fragments were fixed by adding formaldehyde to a final concentration of 3% and were stained with crystal violet. A cell sheet treated with 1 mg/ml of the positive control mAb AK23 was included in each assay to adjust for inter-assay variability. The mean number of particles was determined by counting with Image Pro software, using three sets of digital images captured for each plate. Dissociation scores were calculated from the number of fragments (N) as follows: Dissociation score =((N with mAb-N without mAb)/(N with AK23-N without mAb)) x 100.