EFFECT OF COMPLEMENT AND POLYMORPHONUCLEAR LEUKOCYTE DEPLETION ON EXPERIMENTAL SKIN LESIONS RESEMBLING SYSTEMIC LUPUS ERYTHEMATOSUS P. G. NATALI, M. MOTTOLESE, and M. R. NICOTRA Immunopathologic cutaneous lesions resembling human systemic lupus erythematosus (SLE) can be induced in mice sensitized to ultraviolet (UV)-irradiated DNA following whole body irradiation with UV light. T h e lesions are characterized by the formation of immune complexes at different skin sites. The role played by cellular and humoral mediators in the pathogenesis of this experimental model was investigated. The results obtained suggest that inflammation that follows UV radiation is the major factor responsible for this pathology. Accordingly mice that were rendered neutrophil (PMN) deficient did not manifest skin lesions, and depletion of C3 complement component left them unchanged. I n addition time course studies showed that PMN depletion did not prevent a delayed skin involvement. Thus multiple factors seem to mediate the onset of the immunopathologic changes previously described. Although native DNA is not antigenic in experimental animals, different kinds of physically and From Laboratorio Irnrnunologia Istituto Regina Elena, Rome, Italy. Supported in part by Laboratorio Tecnologia Biornedica, CNR, Italy. Pier Giorgio Natali, M.D.: Istituto Regina Elena; M. Mottolese, M.D.: Istituto Regina Elena; M. R. Nicotra, M.D.: Istituto Regina Elena. Address reprint requests to Dr. Natali, Istituto Regina Elena, l’iale Regina Elena, 291, Rome, Italy. Submitted for publication September 24, 1974; accepted February 14, 1975.
Arthritis and Rheumatism, Vol. IS, No. 6 (November-December 1975)
chemically altered DNA molecules are able to induce a specific immune response (1,2). Native DNA irradiated with ultraviolet (UV) light (UV-DNA) is highly immunogenic (3) and has been used in experimental animal models to elucidate some of the immunologic events leading to kidney (4) and skin (5) involvement in systemic lupus erythematows in humans. Cutaneous lesions in particular could be induced with high frequency in mice with circulating antibodies to UV-DNA following exposure to UV light. This treatment has been shown to be capable of inducing formation of the UV-DNA photoproduct, thymidine dimer (6), and of releasing it from the skin into the blood stream (7). The skin immunopathology closely resembled human lesions in that in vivo fixation of mouse 7globulins and C3 complement component to epidermal cell nuclei and dermal-epidermal junction could be observed together with accumulation of inflammatory cells. In order to investigate the role played by cellular and humoral mediators in the formation of these lesions, the effect of polymorpholeukocyte (PMN) and complement (C3) depletion on the appearance of skin immunopathology was investigated. Results of experiments reported here showed that PMN accumulation in the skin, which follows UV light irradiation, is indispensable for the onset of the cutaneous pathology. C3 depletion prevented only in vivo binding of the complement component to the skin sites where mouse y-globulins were bound to UV-DNA antigen.
NATAL1 E T AL
Fig 1. Effect of nitrogen mustard injection (400 PglIOO g body weight) on PMN number in C57/B1 (0-0) and Balb/C (@-a) mice. PMN numberlmm’ is represented on the ordinate. Each point represents the mean of 5 mice per group. Maximal depletion was reached 7 2 hours after drug injection.
A delayed binding of mouse 7-globulins t o epidermal cell nuclei was observed 24 hours after irradiation even i n the absence of circulating PAIN.
MATERIALS AND METHODS Antigen Preparation and Animal Immunization. Native calf thymus DNA in highly polymerized form (Worthington Biochemical Corporation, Freehold, New Jersey) was irradiated at a concentration of 500 &ml in phosphate (0.01 M ) buffered saline (0.15 M ) with G30T8 germicidal lamps (General Electric, Schenectady, New York) as previously reported (6). After the irradiated DNA was complexed with an equal amount of methylated bovine serum albumin (MBSA), the complex was emulsified in Freund’s adjuvant and injected subcutaneously once a week for at least 4 weeks in either C57/B1 or Balb/C mice strains obtained from a local breeder. Because no difference in response to UV light had been observed in previous studies between the two strains, the experiments reported here were performed on either black (C57/B1) or white (Balb/C) mice. After immunization animals were Med and sera were tested for antibodies to UV-DNA by indirect immunofluorescence as already reported (8.) Animals with antibody titers lower than 1/32 were not used. Polymorph and C3 Depletion. Polymorph (PMN) was depleted by injecting nitrogen mustard, methyl-bis-(Bchloroethyl) amine HCl, “Cloramin” (Simes, Milan, Italy), in two separate intraperitoneal administrations 3 hours apart for a total of 400 &lo0 g of body weight. T h e number of circulating PMN was monitored at different times after red blood cell lysis with Turk-Losung liquid (Merck, Darmstadt, West Germany), and by differential count on blood smears stained with May-Grunwald reagent. Maximal depletion of PMN, 50 to 100 cells/mmY, was reached in both strains of mice at 72 hours and lasted at least 48 hours. C3 was depleted in vivo by injecting purified cobra
venom factor (Cof) supplied by Cordis Laboratories, Miami, Florida. The protein was injected intraperitoneally (20 U/ 100 g of body weight). After a single injection of Cof, the concentration of circulating C3 was estimated by quantitative immunodiffusion using 1 agarose (Seakem, Marine Colloids, Rockland, Maine), 0.05 M EDTA plates, containing an appropriate concentration of a rabbit antimouse C3 specific antiserum (see below). Plates were left to develop at room temperature for 18 hours in a moist chamber. The radial immunodiffusion test was able to detect less than 10% of normal serum C3 concentration. Loss of 90 to 95% of C3 was observed in both strains of mice 24 hours after Cof injection. Low levels of circulating C3 persisted for 2 days, after which a gradual increase of this protein to normal values was noticed. Animal Irradiation. Mice were placed in individual cages as previously described (5). The UV light sources were two G30T8 germicidal lamps. The distance from the light was 50 cm and the irradiation was continued for 7 hours. Before irradiation the horny layer of the epidermis of the ears was stripped off with adhesive tape in order to induce maximal UV-DNA formation i n the epidermal and dermal nuclei. Because in previous experiments most of the pathologic changes following UV light irradiation of the UV-DNA immunized mice were detected in the skin of the ears, all the results reported here were obtained from studies of this skin. Tissue Sampling. Biopsies of the ears were taken either immediately after 7 hours of UV light irradiation or after 24 hours, under ether general anesthesia. All specimens were divided into two parts and processed for routine histology (Bouin’s fixative) or snap frozen at -60°C in Tissue-Tek O.C.T. compound (Ames Company, Elkhart, Indiana) in a mixture of dry ice-acetone. Cryostat sections, 4 p thick, were used in direct and indirect immunofluorescence after being washed for 5 minutes in cold phosphatebuffered saline (PBS). A minimum of three nonconsecutive sections were examined from each biopsy. Reagent Antisera. Antisera to mouse and rabbit 7 S immunoglobulins labeled with fluorescein isothiocyanate were purchased from Hyland Laboratories, Los Angeles, California. The sera appeared monospecific in immunoelectrophoretic analysis and at the dilution used showed an F / P of 4 and a protein concentration of 1.7 mg/ml (antimouse-7-globulin antiserum) and F/P 2.5 with a protein content of 0.8 mg/ml (antirabbit-7-globulin antiserum). Rabbit antiserum to mouse C3 fraction of complement was labeled with fluorescein isothiocyanate by the method of Wood (10). A t the dilution used the serum had a F/P of 4 and a protein concentration of 0.7 mg/ml. Rabbit antiserum to UV-DNA was obtained as previously described (6).
RESULTS Effect of PMN and C3 Depletion on Histologic Changes Induced by U V Light. Figure 1 shows the effect of injection of nitrogen mustard on the number of circulating polymorphs. W h e n 400 pg/IOO g of body weight were administered i n two separate doses,
COMPLEMENT AND PMN LEUKOCYTE DEPLETION
Fig 3. Serum C3 defiletion after injection of Cof (20 p l I 0 0 g body weight) in C 5 7 / B l (0-0) and BalblC (A-A) mice. Each point represents the mean of 4 mice per group. Low levels of circulating C3 lasted 48 hours (black horizontal bar).
Fig 2. a ) Histology of the skin of the ears of PMN-deficient mice after UV radiation. No inflammatory cells are detectable. H+E stain (X 250). b ) Histologic feature of the skin of the ears of C3 depleted mice after UV exposure. Large numbers of PMN infiltrate the upper dermis. Vacuoliration and pyknosis of efiidermal cell nuclei are apparent. H+E stain (X 400).
a sharp decrease in the concentration of PMN was observed both in C57/B1 and Balb/C mice. At 72 hours following drug injection an average of 50 to 100 PMN/mm3 could be detected. T h e low number of neutrophils lasted for at least 48 hours. When normal or UV-DNA immunized animals, which were PMN-depleted, were irradiated according to the experimental conditions that are known to induce a strong inflammatory response, the histologic observation (Figure 2a) of the skin biopsies taken im-
mediately after irradiation showed no PMN accumulation either in the dermis or the epidermis. Few scattered mononuclear cells could be detected in the dermis. T h e epidermal cells showed only minor changes consisting mainly of pyknosis and vacuolization on the side of maximal exposure. These findings contrasted sharply with the effects of in vivo depletion of C3 complement component. Mice injected with Cof showed after 24 hours a reduction of C3 to a concentration less than 10% of normal (Figure 3). When C3-depleted mice were exposed to UV light a normal inflammatory response in the skin was observable. Histologic studies (Figure 2b) showed neutrophils accumulated in large numbers all through the dermis, filling dermal vessels and in close contact with the dermal-epidermal junction. In this instance epidermal cells showed scattered pyknotic nuclei in the basal layer. Incidence and Patterns of in vivo Bound Ig and Complement in the Skin of Nitrogen Mustard- and Cof-Treated Mice. Table 1 summarizes the patterns and incidence of skin fixation of mouse Ig and C3. UV-DNA immunized mice that were UV light irradiated showed maximal incidence of immunopathologic lesions in the skin of the ears immediately after exposure. Direct immunofluorescence studies showed mouse Ig (lOO~o)and C3 (60y0)bound in vivo to epidermal and dermal cell nuclei (Figure 4a). Fixation of the same immune reactants at the dermal-epidermal
NATAL1 E T AL
Table 1. Eflect of PMN and C3 Depletion on Zn Vivo Bound Mouse y-Globulins and C3 in Zmmune Mice Irradiated w i t h UV Light Patterns of Skin Fixation* Dermal Speckled and Reticular Untreated Mouse y-globulin Mouse C3 UV-DNA PMN depleted Mouse 7-globulin Mouse C 3 UV-DNA C3 depleted Mouse y-globulin Mouse C3 UV-DNA
Nuclear 5 / 5 (100) 3 / 5 (60) 515 (100) 015 (0) 015 (0) 5 / 5 (100) 5 / 5 (100) 0 / 5 (0) 515 (100)
*Ear biopsies taken immediately after UV exposure. +Number of positive per number tested (yopositive).
junction (Figure 4b) occurred in 60y0 of the biopsies for mouse Ig and in 40% of those for mouse C3. As expected the antigen UV-DNA was present in all skin biopsies (Figure 4c). Quite striking differences were observed with immune animals that were PMN deficient. UV radiation in these mice, although able to induce the formation of UV-DNA antigen in skin nuclei, did not show in vivo fixation of mouse Ig and C3 either on nuclei or at the dermal-epidermal junction. On the other hand, mice that received Cof and at the time of UV light irradiation had only 10% of the original C3 concentration did not manifest any in vivo binding of the complement component at tissue sites. C3 depletion did not interfere with the amount of mouse anti-UV-DNA antibody that reacted with the homologous antigen, which was induced in all the animals. Nuclear fixation of immunoglobulins was in fact present in 1 0 0 ~ of o the C3-deficient mice. Time Course Study of Skin Lesions in Polymorph-Deficient Mice. I n order to determine whether suppression of immunopathologic lesions in mice that were PMN depleted persisted over a period of time, skin biopsies were sampled in a group of mice both at the end of irradiation and 24 hours later. T h e results of this experiment are summarized in Table 2. Although skin biopsies collected immediately after UV light radiation did not manifest any in vivo binding of either immunoglobulins or C3 (Figure 5a), tissue samples taken in adjacent sites after 24 hours demonstrated in all cases mouse 7-globulins fixed to epidermal cell nuclei (Figure 5b). Dermal-epidermal
Fig 4. a ) Direct iinniiinofliiorescence showing in vivo binding of mouse y-globulins at the epidermal nuclei in UV-immunized, UV light-irradiated mice ( X 250). b ) Reticular distribution of y-globulins in the upper dermis and at the dermal-epidermal junction in the same animals ( X 250). c ) Zndirect immunopuorescence to detect U V - D N A in the epidermis of UV light-irradiated mice. Epidermal and dermal nuclei show bright fluorescence after reaction w i t h a rabbit anti-UV-DNA antiserum (X 250).
junction stain was very seldomly observed. T h e extent of the lesions was generally limited to some areas of the epidermis with no appreciable involvement of the deep dermal structures. No C3 was detectable with the same distribution. These findings strongly suggested that, although neutrophils were essential to the onset of the early lesion, they were completely unrelated to the appearance of the immunopathology observed 24 hours after irradiation. In fact the number of peripheral neutrophils in most cases was further decreased or unchanged. Titers of circulating antibodies to UV-DNA also seemed not to play any relevant role in such phenomena.
DISCUSSION Formation of immune complexes either in circulation or at tissue sites appears to be one of the most
COMPLEMENT AND PMN LEUKOCYTE DEPLETION
of inflammatory cells together with in vivo deposition of mouse y-globulins and C3 to different skin sites. Because complement activation and neutrophil accumulation are two important phenomena in immune inflammation, the effect of their suppression in vivo in this animal model was investigated. Depletion of the third complement component achieved by Cof did not interfere with the binding of circulating mouse antiUV-DNA antibodies to the homologous antigen in the skin or with accumulation of neutrophils. This finding suggests that complement activation plays only a secondary role in eliciting these skin lesions. UV light has the ability to induce increased vascular permeability and PMN accumulation independently from the presence of a specific immune reaction at the skin level (5). Mice that were PMN deficient, on the contrary, completely lacked any immunopathologic manifestation. No in vivo fixation of mouse 7-globulins to the skin was observable after UV light exposure. This result can also be explained by the fact that UV light is a strong inflammatory stimulus mainly mediated by polymorpholeukocytes (12). T h e accumulation of these cells together with the increase in vascular permeability may be induced by UV light through various nonimmune mechanisms involving changes of cell membrane permeability (13), liberation of vasoactive amines from tissue reservoirs (14) and disruption of lysosomes of pre-existing and accumulating cells (15,16) with liberation of enzymatic activities and leukotactic factors (17). Although nitrogen mustard does not selectively affect leukocyte subclasses, the clrug does subtract the cell population that is primarily engaged in the UV light inflammatory response by depleting most of the circulating polymorphs. Such response is likely to allow circulating anti-UV-DNA antibodies to react with newly formed UV-DNA antigen, adding a further in-
Fig 5. a ) Lack of any in vivo binding of mouse y-globulins in biopsies collected immediately after U V radiation in PMN-deficient mice (X 250). Autofluorescence of the upper epidermis is present. b ) Ear biopsies taken from adjacent sites 24 hours after UV light show mouse y-globulins fixed to epidermal nuclei ( X 250).
common mechanisms that bring about various cellular alterations such as necrosis, proliferation, inflammatory cell accumulation, and increased vascular permeability (1 1). T h e skin lesions previously described (5) were shown to be mediated presumably by the formation of immune complexes between UV-DNA antigen and homologous antibody. Thus the simultaneous presence of both immune reagents was indispensable for the appearance of cutaneous pathology. This appearance was characterized by a n accumulation
Table 2. Time-Course of Skin Lesions in PMN-Depleted Mice after UV Exposure
64 64 64
150 200 200
*Reciprocal of serum antibody titerto UV-DNA. iIn vivo binding to epidermal cell nuclei.
+ + + + +
60 100 150
NATAL1 E T AL
flammatory insult of immune nature. I t should be stressed that the nonselective activity of the drug does not exclude a possible participation of other humoral and/or cell-mediated mechanisms in the onset of this pathology, although at the high levels of exciting radiation delivered to the animals polymorph accumulation should be of pathogenetic importance. Despite the low number of circulating PMN, binding of anti-UV-DNA antibodies to cell nuclei was again observed in the nitrogen mustard-treated mice 24 hours after irradiation. This finding agrees with previous studies that demonstrated that the increase in vascular permeability following moderate doses of UV light radiation is characterized by a diphasic response (17). T h e first response is mediated through liberation of amines such as histamine (guinea pig) and 5-hydroxytryptamine (rats) and is accompanied by emigration of leukocytes; the second occurs 24 hours after the UV stimulus and is completely independent of the presence of inflammatory cells. T h e mediators of this delayed permeability increase remain unknown. T h e failure of C3 to bind at 24 hours to the epidermal cell nuclei is not yet understood. Although nitrogen mustard did not reduce C3 serum levels, an effect of the drug on other serum components that participate in C3 activation cannot be excluded. T h e data presented here suggest that this experimental model may provide a tool for gaining more insight into the mechanism involved in the autoimmune manifestations of SLE. T h e same model may also help to explain the role that different humoral and cellular mediators have in this cutaneous pathology. Such knowledge will be valuable in guiding the therapeutic approach to this disease.
REFERENCES 1 . Plescia OJ, Braun W, Palczuk NC: Production of antibodies to denatured deoxyribonucleic acid (DNA). Proc Natl Acad Sci USA 52:279-285, 1964 2. Van Vunakis H, Seaman E, Kahan L, et al: Formation of an adduct with tris (hydroxymethyl) amino ethane during the photoxidation of deoxyribonucleic and guanidine derivatives. Biochemistry 5:3986-3991, 1966
3. Natali PG, T a n EM: Antibodies to photoproducts of DNA irradiated with ultraviolet light. Fed Proc 28:696, 1969 4. Natali PG, T a n EM: Experimental renal disease induced by DNA anti-DNA immune complexes. J Clin Invest 51:345-355, 1972 5. Natali PG, T a n EM: Experimental skin lesions in mice resembling systemic lupus erythematosus. Arthritis Rheum 16:579-589, 1973 6. Natali PG, T a n EM: Immunological detection of thymidine photoproduct formation in vivo. Radiat Res 46:506-5 18, 1971 7. T a n EM: Production of potentially antigenic DNA in cells. Immunopathology, Sixth International Symposium. Edited by PA Mieschen. Basel, Benro Schwabe CO, 1971, p p 346-349 8. T a n EM: Antibodies to deoxyribonucleic acid irradiated with ulraviolet light. Detection by precipitins and immunofluorescence. Science 161:1353-1354, 1968 9. Mardiney MK, Muller-Eberhard HJ: Mouse BIC-globulin. Production of antiserum and characterization in the complement reaction. J Immunol 94:877-882, 1965 10. Wood BT, Thompson SH, Goldstein G : Fluorescent antibody staining. 111. Preparation of fluorescein-isotliiocyanate labelled antibodies. J Immunol 95:225-229, 1965 11. Dixon FJ: T h e role of antigen-antibody complexes in disease. Harvey Lect 58:21-52, 1962-1963 12. Logan G, Wilhelm DL: T h e inflammatory reaction in ultraviolet injury. Br J Exp Pathol 46:286-299, 1966 13. Kinaldi RA: T h e effect of ultraviolet radiation on the hydration of amoeba proteus as determined by contractile vacuolar function. Exp Cell Res 18:62-69, 1959 14. Logan G, Wilhelm DL: Vascular permeability changes in inflammation. I. T h e role of endogeneous permeability factors in ultraviolet injury. Br J Exp Pathol 47:300314, 1966 15. Johnson BE: Ultraviolet radiation and lysosomes in skin. Nature 219: 1258-1259, 1968 16. Desai ID, Savant PL, Tappel AL: Peroxidative and radiation damage to isolated lysosomes. Biochem Biophys Acta 86:277-285, 1964 17. Janoff A, Zweifach BW: Production of inflammatory changes in the microcirculation by cationic proteins extracted from lysosomes. J Exp Med 120:747-764, 1964 18. Logan G, Wilhelm DL: Ultraviolet injury as an experimental model of the inflammatory reaction. Nature 198:968-969, 1963