• smooth muscle antibodies
• antinuclear antibodies
• Hep2 cells
• type 1 autoimmune hepatitis
• actin specific antibodies

We read with great interest the article by Silvestrini et al.¹ In November 1999, patient 1 (table 1) presented with malaise, arthralgia, thrombocytopaenia, low complement values, abnormal liver function tests, a strong homogeneous antinuclear antibody (ANA) result, and smooth muscle antibodies. We demonstrated that this patient's smooth muscle antibodies were specific for actin using Hep2 cells, a result that is suggestive of type 1 autoimmune hepatitis, and which was later confirmed by liver biopsy. Since then, we have tried to determine the target specificity of smooth muscle antibodies in our laboratory, which serves three hospitals and carries out over 10 000 autoantibody tests each year.

Unlike the confirmatory tests described in the literature,¹ Hep2 cells are routinely available. The human epithelial cell monolayer has been regarded as unreliable for the detection of actin specific antibodies, because the staining for actin varies from speckles, as a result of truncated cables, to randomly distributed filaments.² However, we have now used Hep2 cells (fig 1) to confirm the actin specificity in 18 patient samples (table 1) following characteristic staining on rat liver, kidney, and stomach. We have had no difficulty in distinguishing actin specific antibodies from other cytoskeletal antibodies. We have recently started using a composite block for routine autoantibody screening, consisting of rodent kidney and stomach, and primate liver, together with human epithelial cells with good actin expression.

We are currently assessing the proportion of smooth muscle antibodies that are actin specific, and its predictive value for autoimmune hepatitis. We also describe three patients with biopsy confirmed type 1 autoimmune hepatitis (AIH) and antibodies directed against double stranded DNA, but no actin specific antibodies (table 2).

Although highly specific for type 1 AIH, actin specific antibodies have been described in primary biliary cirrhosis (PBC), alcoholic liver disease, connective tissue disorders, and healthy people.^3,4. In our cohort, patients 8, 12, and 13 have thyroid disease. Patient 13 also had anti-Hu antibody (also known as ANNA-1, anti neuronal nuclear antibody 1), dementia, and oligoclonal bands in the cerebrospinal fluid, suggestive of a paraneoplastic disorder. Patient 6 has M2 antimitochondrial antibodies, suggestive of a PBC/AIH overlap. Patients 1–5, 7, and 17 (table 1) have had liver biopsies, the results of which were compatible with type 1 AIH.

We feel that the use of Hep2 cells is an easily applicable confirmatory test for actin antibodies. However the sensitivity and specificity of actin antibodies for type 1 autoimmune hepatitis should still be regarded with some caution.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Characteristic staining of microfilaments in human epithelial cells by an anti-actin antibody.