Sexual dimorphism in innate immunity

To establish whether variation in innate immunity, as measured by the level of tumor necrosis factor (TNF) in lipopolysaccharide (LPS)-stimulated whole-blood culture, is related to sex or HLA.

Normal volunteers (72 women, 159 men) completed questionnaires and donated peripheral blood specimens. Blood samples were exposed to LPS in a 4-hour in vitro culture, and supernatants were then tested by sandwich-type immunoassay measuring TNF levels. Statistical techniques included multivariate analysis and maximal-likelihood modeling of allelic effects.

Both male and female groups showed substantial within-group variation (coefficient of variation 59.1% for women, 40.3% for men). However, the mean +/- SD LPS-stimulated TNF level in the female group was nearly 30% lower than in the male group (1,556+/-919 pg/ml versus 2,203+/-889 pg/ml; P < 0.0001, unadjusted for covariates). Sex was independent of any microsatellite marker allele of TNF (covariate-adjusted increment of 785 pg/ml from female to male sex; P < 0.0001). In multivariate modeling of the female group, the LPS-stimulated TNF level was not independently influenced by menstrual cycle phase, oral contraceptive use, or plasma estradiol level. Allelic modeling showed that significant TNFab microsatellite allelic effects existed (P = 0.002 versus model omitting allelic effects). The female group showed a significantly downward deviation from mean TNF level with TNFa4b5 (-903 pg/ml deviation from the overall mean) and an upward deviation with TNFa10b4 (598 pg/ml). The male group showed significantly higher-than-mean levels with TNFa1b5 (909 pg/ml), TNFa5b7 (1,191 pg/ml), and TNFa6b5 (332 pg/ml). Thus, the two sex groups differed in which of their TNFab marker alleles showed significant deviations from the overall mean.

Female subjects have a nearly 30% lower innate immune response, stemming largely from influence independent of the HLA-region TNF locus and without further independent variation stemming from plasma estrogen level.