We have restudied two kindreds that formed the basis of the original r
eport of autosomal recessive chronic granulomatous disease (CGD) assoc
iated with leukocyte glutathione peroxidase deficiency. Case 1 from th
e original study and the surviving brother of the originally reported
case 2 both have severe CGD, with no detectable respiratory burst acti
vity in purified intact neutrophils. However, their leukocytes exhibit
normal glutathione peroxidase enzyme activity and gene expression. Ex
amination of phagocyte nicotinamide adenine dinucleotide phosphate (NA
DPH)-oxidase components known to be defective in CGD reveals no detect
able cytochrome b(558) nor any membrane activity in a cell-free NADPH
oxidase assay system. Molecular analysis of the genes encoding cytochr
ome b(558) subunits shows, in case 1, a C --> T substitution at nucleo
tide 688 of the gene encoding the gp91-phox subunit of cytochrome b(55
8), resulting in termination signal in place of Arginine-226. Levels o
f gp91-phox mRNA are markedly decreased despite normal levels of gene
transcription, indicating a post-transcriptional effect of the nonsens
e mutation on mRNA processing or stability. The X-linked form of CGD d
eveloped in this cytogenetically normal female due to the uniform inac
tivation of the normal X chromosome in her granulocytes, indicated by
the expression in her granulocyte mRNA of only one allele of a glucose
-6-phosphate dehydrogenase polymorphisms for which she is heterozygous
in genomic DNA. Case 2 (of the present study) has distinct mutations
in each allele of the p22-phox gene. Thus, the CGD phenotype in both k
indreds now appears to be caused not by leukocyte glutathione peroxida
se deficiency, but rather by mutations of the same subunits of phagocy
te NADPH oxidase that are involved in all other cases of cytochrome b-
negative CGD. Glutathione peroxidase deficiency should no longer be co
nsidered a possible cause of the CGD phenotype. (C) 1994 by The Americ
an Society of Hematology.