Do tissues fixed in a non-cross-linking fixative require a dedicated formalin-free processor?

We evaluate the consequences of processing alcohol-fixed tissue in a processor previously used for formalin-fixed tissue. Biospecimens fixed in PAXgene Tissue Fixative were cut into three pieces then processed in a flushed tissue processor previously used for formalin-fixed, paraffin-embedded (FFPE) blocks (neutral buffered formalin [NBF](+ve)), a formalin-free system (NBF(-ve)), or left unprocessed. Histomorphology and immunohistochemistry were compared using hematoxylin/eosin staining and antibodies for MLH-1, Ki-67, and CK-7. Nucleic acid was extracted using the PAXgene Tissue RNA/DNA kits and an FFPE RNA extraction kit. RNA integrity was assessed using RNA integrity number (RIN), reverse transcription polymerase chain reaction (RT-PCR) (four amplicons), and quantitative RT-PCR (three genes). For DNA, multiplex PCR, quantitative PCR, DNA integrity number, and gel electrophoresis were used. Compared with NBF(-ve), RNA from NBF(+ve) blocks had 88% lower yield and poorer purity; average RIN reduced from 5.0 to 3.8, amplicon length was 408 base pairs shorter, and Cq numbers were 1.9-2.4 higher. Using the FFPE extraction kit rescued yield and purity, but RIN further declined by 1.1 units. Differences between NBF(+ve) and NBF(-ve) in respect of DNA, histomorphology, and immunohistochemistry were either non-existent or small in magnitude. Formalin contamination of a tissue processor and its reagents therefore critically reduce RNA yield and integrity. We discuss the available options users can adopt to ameliorate this problem.