Linear IgA bullous dermatosis in a patient with acute lymphocytic leukemia: possible involvement of granulocyte colony-stimulating factor

ARTICLE

Linear IgA bullous dermatosis (LABD) is an acquired autoimmune blistering disease characterized by linear IgA deposition at the basement membrane zone and neutrophilic microabscesses in the papillary dermis. Although the pathogenesis of LABD is still unknown, there have been many reports of its association with drug exposure [1-25] as well as internal malignancies [26, 27]. Granulocyte colony-stimulating factor (G-CSF) is a hematopoietic factor that promotes the proliferation and differentiation of neutrophils [28], and has been implicated in the induction of Sweet's syndrome [29, 30] and bullous pyoderma gangrenosum [31]. G-CSF, however, has never been reported in association with LABD. We describe a case of LABD in a patient with acute lymphocytic leukemia (ALL) following a drug eruption and administration of G-CSF.

Case report

A 64-year-old woman (partly described elsewhere [32]) had a low-grade fever of unknown origin for 2 months. She was admitted to a hospital on May 17, 1995, because of persistent leukopenia. Because acute leukemia was suspected, she was referred to our hospital on June 16 after the administration of 250 µg of G-CSF daily for 1 week. Her past medical history was otherwise unremarkable. Antibiotic therapy for presumptive infection was initiated with a combination of imipenem cilastatin sodium (IPM) and amikacin sulfate (AMK) on June 19. On day 10 of the antibiotic treatment, discrete erythematous macules were noted on the trunk. Lesions on the trunk extended to the extremities; the antibiotics were stopped on June 28 and the eruptions disappeared within 7 days despite continuation of G-CSF. Even after the discontinuation of the antibiotic treatment, diclophenac was frequently utilized because the patient had a spiking temperature up to 39° C. Enlarged cervical lymph nodes were palpable but not tender. Histopathological findings of the biopsy specimen from one of the lymph nodes were consistent with tuberculosis.

A combination therapy of rifampicin (RFP), isoniazid (INH), and streptomycin (SM) was started on July 3 and was continued for 6 months. The administration of G-CSF was only interrupted for 1 week after her white blood cell count reached approximately 4 x 10⁹/L; however, because the patient again developed leukopenia, G-CSF 100 µg daily restarted. Bullous eruptions appeared on the trunk on July 13. Physical examination revealed approximately 10 tense blisters with slight erythema on the trunk (Fig. 1) and upper limbs. No oral or genital involvement was noted. Regardless of the onset of bullous eruption, diclophenac was given when she had high-grade fever during this period.

Relevant laboratory values were as follows: white blood cell count 4.9 x 10⁹/l, with 0.01 myelocytes, 0.81 neutrophils, 0.17 lymphocytes, and 0.01 atypical lymphocytes; platelets 126 x 10⁹/l; serum protein, 70 g/l; serum albumin, 30 g/l; erythrocyte sedimentation rate, 82 mm/h; C-reactive protein elevated at 26 mg/l, and negative blood cultures. Liver function studies, chest X-ray, and CT scan of the abdomen were within normal limits. A bone marrow biopsy revealed a normocellular marrow. Approximately 80% of nucleated cells were blasts with cleft and prominent nucleoli. Myeloperoxidase was negative in most of the blasts. The patient was diagnosed as having ALL.

A skin biopsy specimen from the bullous lesion revealed a subepidermal bulla with an inflammatory infiltrate consisting mainly of neutrophils and eosinophils in the papillary dermis (Fig. 2). Direct immunofluorescence studies of the lesion revealed linear deposition of IgA along the basement membrane zone (Fig. 3), but no deposit of IgG, IgM or complement. Indirect immunofluorescence, using 1 mol/L NaCl-split skin as a substrate, demonstrated that the serum obtained during the development of bullous lesions showed weak IgA deposition localized to the epidermal side; no circulating IgA antibodies were detected in the serum obtained after the resolution of the bullous eruption. Immunoblotting on serial serum samples gave negative results. Based on the histopathological and immunofluorescence findings, a diagnosis of LABD was made. Approximately 4 weeks after the discontinuation of G-CSF treatment, the lesions resolved spontaneously, leaving pigmentation. Treatment with G-CSF, 100 µg daily, was again initiated in view of the beneficial effect on her severe hematological condition; but no recurrence was noted.

Rechallenge with IPM reproduced the erythematous macules on the trunk within 48 hrs. Therefore, the erythematous macular lesions which were observed before the appearance of bullous lesions in this patient were diagnosed as a drug eruption due to IPM. The patient has avoided IPM, but not G-CSF, anti-tuberculosis drugs and diclophenac, since then and has had no further recurrence in the past 6 months.

Discussion

Drug-induced LABD differs in several respects from idiopathic LABD, as suggested by Kuechle et al. [15]: cutaneous symptoms are transient with a lack of mucosal or conjunctival lesions and resolve rapidly when the implicated drug is discontinued; immune deposits of IgA at the basement membrane zone disappears from the skin once the lesions have resolved; and most patients lack circulating IgA antibodies. The clinical course of our patient is consistent with drug-induced LABD, because oral involvement was not found and discontinuation of G-CSF resulted in complete clearance within 4 weeks. The only difference was the presence of circulating IgA localized to the epidermal side of salt-spilt human skin in our patient. In contrast, Kuechle et al. [15] reported that only one of six patients with drug-induced LABD had circulating antibodies but they were localized to the dermal side. Nevertheless, serum obtained after resolution of the eruption did not show IgA deposition, a finding suggestive of drug-induced LABD.

A number of drugs have been reported to induce LABD: the frequently reported drugs in association with LABD include diclophenac [1, 3, 14, 23], sulphamethoxazole [23], vigabatrin [23], penicillins [2, 13, 25], glibenclamide [4], lithium carbonate [5], cefamandole nafate [6], iodine [7], vancomycin [8, 11, 12, 15, 17-21], captopril [9, 15], a combination of interleukin-2 (IL-2)/interferon-gamma (IFN-gamma) [10], or IL-2 alone [22], polychemotherapy [11], somatostatin [15], phenytoin [15, 24], and amiodarone [16]. One should be cautious in interpreting such association however, because definite proof of association is lacking in most patients: a diagnosis of drug-induced LABD was made exclusively on the basis of complete resolution of the eruption after discontinuation of the suspected drug without confirmation by recurrence of symptoms after reinstitution of the drugs, except for several cases [1, 4, 5, 8] in which rechallenge reproduced the eruptions within hours to days. Probably for ethical reasons rechallenge would not have been attempted in many cases. Indeed, we would have concluded that the LABD can be attributed to diclophenac but not to G-CSF based on the increased number of reported cases, if our patient had avoided taking diclophenac after complete resolution of the eruption. However, this was not the case.

The interval between the onset of LABD and commencement of the drug more favors G-CSF rather than diclophenac as the probable cause of LABD. In view of its ability to stimulate the proliferation, differentiation, and chemotaxis of neutrophils, it is logical to consider that G-CSF would play an important role in the development of LABD by activating neutrophils although not per se sufficient for the outbreak of the disease. Support for the role of G-CSF in the development of LABD is found in previous suggestion by Hendrix et al. [33] that IgA deposits in LABD may act as specific ligands to mediated neutrophil inflammatory functions and that neutrophil-tissue IgA interactions may be enhanced by G-CSF. However, recurrence of LABD was not induced by the reinstitution of G-CSF after the complete resolution in this patient. This is not easily reconciled with the notion that G-CSF was directly involved in the development of the LABD unless one assumes the presence of precipitating factors that can synergistically enhance and accelerate the outbreak of the disease in our patient treated with G-CSF. This multifactorial process resulting in the development of the LABD could explain why a complete resolution of the LABD was delayed for 4 weeks despite discontinuation of G-CSF.

There are at least three plausible explanations as to why the reinstitution did not reproduce the bullous eruptions: firstly it might be that cytokines, such as IL-2 and IFN-gamma, endogenously produced during the development of a drug eruption provide a favorable milieu for the onset of G-CSF-induced LABD. In support of this possibility, previous documents described patients with renal cell carcinoma in whom LABD appeared while undergoing immunotherapy with an IL-2/IFN-gamma combination [10] or IL-2 alone [22]. These observations indicate that overproduction of various cytokines from activated lymphocytes, such as IL-2, tumor necrosis factor, and IFN-gamma may have a prerequisite for the subsequent outbreak of the LABD. If this view is true, then the reported association between LABD and malignancies including lymphoproliferative disorders [26, 27] could now be interpreted as an indication that the combination of different cytokines produced endogenously in these conditions may be responsible for the development of LABD even in patients without cytokine therapy, because it has been demonstrated that a variety of cytokines such as interleukin-1 (IL-1), interleukin-8 (IL-8), G-CSF, and macrophage colony-stimulating factor (GM-CSF) were endogenously secreted in a dysregulated fashion in patients with myelogenous disorders [34]. Secondly, because G-CSF has been shown to stimulate neoplastic cells [35], it is possible that a variety of cytokines endogenously secreted by leukemic cells, in addition to G-CSF administered, may have contributed to the development of the bullous lesions in our patient. Thirdly, it is also likely that increased expression of LABD-related, sequestered autoantigens can be induced as a consequence of epidermal damage by activated T cells during a drug eruption, and thereby causing autoantibody formation and their subsequent deposits along the basement membrane zone in our patient.

We suggest that patients with LABD will need special attention with respect to the type of cytokines or combination of cytokines given as therapeutic modalities.

REFERENCES

1. Gabrielsen TØ, Staerfelt F, Thune PO. Drug-induced bullous dermatosis with linear IgA deposits along the basement membrane. Acta Derm Venererol (Stockh) 1981; 61: 439-41.

2. Boffety B, Sohier J, Venencie PY, Laumont T, Blanchet-Barden C, Puissant A. Dermatoses bulleuses à IgA linéaire. Journées Dermatologiques de Paris 1981: 53-53a.

3. Valsecchi R, Serra M, Tornaghi A, Cainelli T. Dermatosi bollosa da farmaco a IgA lineari. G Ital Dermatol Venereol 1982; 117: 221-3.

4. Väätäinen N, Fräki JE, Hyvönen M, Neittaanmäki H. Purpura with a linear epidermo-dermal deposition of IgA. Acta Derm Venereol (Stockh) 1983; 63: 169-70.

5. McWhirter JD, Hashimoto K, Fayne S, Ito K. Linear IgA bullous dermatosis related to lithium carbonate (Letter). Arch Dermatol 1987; 123: 1120-2.

6. Argenyi ZB, Bergfeld WF, Valenzuela R, McMahon JT, Taylor JS. Adult linear IgA disease associated with an erythema multiforme-like drug reaction. Cleve Clin J Med 1987; 54: 445-50.

7. Lemarchand-Venencie F, Vigouroux F, Blanc F, Pawin H, Pons A, Bonvalet D, et al. Dermatose bulleuse à IgA linéaire au décours immédiat d'une urographie intra-veineuse. Ann Dermatol Venereol 1987; 114: 1423-5.

8. Baden LA, Apovian C, Imber MJ, Dover JS. Vancomycin-induced linear IgA bullous dermatosis (Letter). Arch Dermatol 1988; 124: 1186-8.

9. Klein LE, Shmunes E, Carter JB, Walsh MY. Linear IgA bullous dermatosis related to captopril treatment. Cutis 1989; 44: 393-6.

10. Guillaume JC, Escudier B, Espagne E, Roujeau JC, Prost C, Domart P, et al. Dermatose bulleuse avec dépôts linéaires d'IgA le long de la membrane basale au cours d'un traitement par l'interféron gamma et l'interleukine-2. Ann Dermatol Venereol 1990; 117: 899-902.

11. Espagne E, Prost C, Chosidow O, Saïag Ph, Revuz J, Roujeau JC. Dermatoses à IgA linéaire médicamenteuses: à propos de trois cas. Ann Dermatol Venereol 1990; 117: 898-9.

12. Carpenter S, Berg D, Sidhu-Malik N, Hall III RP, Rico MJ. Vancomycin-associated linear IgA dermatosis. J Am Acad Dermatol 1992; 26: 45-8.

13. Cambemale P, Gavaud C, Cozzani E, Nicolas JF, Guennoc B, Dusseau JY. Dermatose à IgA linéaire(dial) induite par Pénicilline G. Ann Dermatol Venereol 1993; 120: 847-8.

14. Collier PM, Wojnarowska F. Drug-induced linear immunoglobulin A disease. Clin Dermatol 1993; 121: 529-33.

15. Kuechle MK, Stegemeir E, Maynard B, Gibson LE, Leiferman KM, Peters MS. Drug-induced linear IgA bullous dermatosis: report of six cases and review of the literature. J Am Acad Dermatol 1994; 30: 187-92.

16. Primka III EJ, Liranzo MO, Bergfeld WF, Dijkstra JW. Amiodarone-induced linear IgA disease. J Am Acad Dermatol 1994; 31: 809-11.

17. Piketty C, Meeus F, Nochy D, Poux JM, Jacquot C, Bariety J. Linear IgA dermatosis related to vancomycin. Br J Dermatol 1994; 130: 130-1.

18. Richards S, Hall S, Yokel B, Whitmore SE. A bullous eruption in an elderly woman: vancomycin-associated linear IgA dermatosis. Arch Dermatol 1995; 131: 1447-52.

19. Geissmann C, Beylot-Barry M, Doutre MS, Beylot C. Drug-induced linear IgA bullous dermatosis. J Am Acad Dermatol 1995; 32: 296.

20. Bitman LM, Grossman ME. Bullous drug eruption treated with amputation: a challenging case of vancomycin-induced linear IgA disease. Arch Dermatol 1996; 132: 1289-90.

21. Whitworth JM, Thomas I, Peltz SA, Sullivan BC, Wolf AH, Cytryn AS. Vancomycin-induced linear IgA bullous dermatosis (LABD). J Am Acad Dermatol 1996; 34: 890-1.

22. Tranvan A, Pezen DS, Medenica M, Michelson GC, Vogelzang N, Soltani KM. Interleukin-2 associated linear IgA bullous dermatosis. J Am Acad Dermatol 1996; 35: 865-7.

23. Paul C, Wolkenstein P, Prost C, Caux F, Rostoker G, Heller M, et al. Drug-induced linear IgA disease: target antigens are heterogeneous. Br J Dermatol 1997; 136: 406-11.

24. Acostamadiedo JM, Perniciaro C, Rogers RS. Phenytoin-induced linear IgA bullous disease. J Am Acad Dermatol 1998; 38: 352-6.

25. Wakelin SH, Allen J, Zhou S, Wojnarowska F. Drug-induced linear IgA disease with antibodies to collagen VII. Br J Dermatol 1998; 138: 310-4.

26. McEvoy MT, Connolly SM. Linear IgA dermatosis: association with malignancy. J Am Acad Dermatol 1991; 22: 59-63.

27. Godfrey K, Wojnarowska F, Leonard J. Linear IgA disease of adults: association with lymphoproliferative malignancy and possible role of other triggering factors. Br J Dermatol 1991; 123: 447-52.

28. Asnis LA, Gaspari AA. Cutaneous reactions to recombinant cytokine therapy. J Am Acad Dermatol 1995; 33: 393-410.

29. Richard MA, Grob JJ, Laurans R, Hess S, Brunet P, Stoppa AM, et al. Sweet's syndrome induced by granulocyte colony-stimulating factor in a woman with congenital neutropenia. J Am Acad Dermatol 1996; 35: 629-31.

30. Prevost-Blank PL, Shwayder TA. Sweet's syndrome secondary to granulocyte colony-stimulating factor. J Am Acad Dermatol 1996; 35: 995-7.

31. Ross HJ, Moy LA, Kaplan R, Figlin RA. Bullous pyoderma gangrenosum after granulocyte colony-stimulating factor treatment. Cancer 1991; 68: 441-3.

32. Muneyuki M, Kokaji T, Kano Y, Shiohara T. Drug-induced linear IgA bullous dermatosis (in Japanese). Rhinsho derma (Tokyo) 1998; 40: 99-102.

33. Hendrix JD, Mangum KL, Zone JJ, Gammon WR. Cutaneous IgA deposits in bullous diseases function as ligands to mediate adherence of activated neutrophils. J Invest Dermatol 1990; 94: 667-72.

34. Negrin RS, Haeuber DH, Nagler A, Olds LC, Donlon T, Souza LM, et al. Treatment of myelodysplastic syndromes with recombinant human granulocyte colony-stimulating factor: a phaseI-II trial. Ann Intern Med 1989; 110: 976-84.

35. Tachibana M, Miyakawa A, Uchida A, Murai M, Eguchi K, Nakamura K, et al. Granulocyte colony-stimulating factor receptor expression on human transitional cell carcinoma of the bladder. Br J Cancer 1997; 75: 1489-96.