IENF-counting by a new morphometric modification is reproducible and diagnostically sensitive and can easily be adopted in any laboratory familiar with the basic immunohistochemical methodology. The method is less dependent on costly technical support systems and seems to be less time consuming when compared with conventional methods for IENF-counting.

We studied the expression of antioxidant enzymes (AOEs) and related proteins manganese superoxide dismutase (MnSOD), thioredoxin (Trx), thioredoxin reductase (TrxR), and the catalytic (GLCL‐c) and regulatory (GLCL‐r) subunits of glutamate cysteine ligase (γ‐glutamylcysteinesynthetase) in 433 astrocytomas. Expression of MnSOD was found in 91%, Trx in 46%, TrxR in 66%, GLCL‐c 73% and GLCL‐r in 89% of the cases. Diffuse astro‐cytomas showed more intense staining for Trx (p= 0.002), TrxR (p=0.004), GLCL‐c (p=0.001), GLCL‐r (p=0.04) and MnSOD (p=0.01) than pilocytic astrocytomas. Within diffuse astrocytomas only Trx (p= 0.0001) and TrxR (p=0.04) significantly associated with increased malignancy grade. Necrotic tumors were more often immunopositive for Trx (p=0.001) and TrxR (p=0.02) and AOE expression was generally higher in mitotically active tumors. Expression of Trx and lack of MnSOD expression was associated with a worse prognosis in diffuse astrocytomas. None of the AOEs had any prognostic value in pilocytic grade I astrocytomas. Familial astrocytomas, which included 23 of the cases studied, did not differ in their expression of MnSOD from sporadic ones. The results show that MnSOD and Trx may influence the biological behaviour of astrocytomas, possibly by modulating cell proliferation and necrosis in these tumors.

We introduce a modification of the tissue microarray technique in which several frozen brain tissue specimens are collected to a single frozen brain array block. In the present application, we use it for the detection of neuronal paraneoplastic anti-Hu autoantibodies. Representative samples from 15 different brain regions were collected according to a standard neuropathological autopsy protocol. Cryostat sections from each block were cut and conventionally stained. From representative areas, cylinder tissue samples from each specimen were punched and then arrayed into a recipient array block. Using the cryostat sections of this brain array, autoantibodies from seven anti-Hu-positive patient sera (confirmed by immunoblotting) were screened by immunohistochemistry. Neuronal architecture was well preserved and immunohistochemical staining was comparable to that of conventional cryostat sections. Because of the variable staining pattern in different brain areas, two anti-Hu-positive sera could be detected immunohistochemically by the one brain array. With the present array technique, it is possible to characterize the variable staining patterns of neuronal paraneoplastic autoantibodies in different locations of the human brain. The frozen brain array also allows the detection of RNA and DNA targets involved in neurological diseases.