Abstract
IMPORTANCE: Schizophrenia (SZ) is characterized by deficits in visual-spatial working memory, a function dependent on a distributed cortical network that includes nodes in the primary visual (V1), posterior parietal (PPC), and dorsolateral prefrontal (DLPFC) cortices. The connections across these regions involve excitatory synapses on the dendritic spines of pyramidal neurons in layer 3 (L3). OBJECTIVE: To assess whether SZ is associated with regional differences in the magnitude or nature of L3 spine alterations. DESIGN, SETTING, AND PARTICIPANTS: This case-control study examined brain tissue from 20 individuals with SZ and 20 matched unaffected comparison (UC) individuals, obtained by the Allegheny County Office of the Medical Examiner (Pittsburgh, Pennsylvania). Dendritic spines were labeled using fluorescent phalloidin (for F-actin) and immunolabeling for spinophilin and imaged by confocal microscopy. EXPOSURE: Schizophrenia. MAIN OUTCOMES AND MEASURES: Primary outcomes were between-group differences in L3 dendritic spine density and size across V1, PPC, and DLPFC. Secondary outcomes included spine fluorescence intensities of phalloidin and spinophilin. RESULTS: Forty individuals were studied (20 UC: 14 [70%] male and 6 [30%] female; mean [SD] age, 47.7 [9.6] years; 20 SZ: 14 [70%] male and 6 [30%] female; mean [SD] age, 45.6 [9.5] years). Dendritic spine density reductions in individuals with SZ varied by spine size across regions, with lower densities of small spines in V1 (-18%; 95% CI, -31% to -5%; P = .009), medium spines in PPC (-16%; 95% CI, -28% to -4%; P = .01) and DLPFC (-13%; 95% CI, -21% to -4%; P = .009), and large spines in PPC (-38%; 95% CI, -58% to -17%; P < .001) and DLPFC (-30%; 95% CI, -50% to -11%; P = .004). Phalloidin fluorescence was lower in small (-9.5%; 95% CI, -17% to -1%; P = .04) and medium (-9.8%; 95% CI, -18% to -1%; P = .04) V1 spines and higher in large DLPFC spines (9.5%; 95% CI, 0.4% to 19%; P = .049). Spinophilin fluorescence was lower across all spine sizes and regions (a range from -13%; 95% CI, -24% to -2%, to -34%; 95% CI, -46% to -21%; P values ranging .02 to <.001). CONCLUSIONS AND RELEVANCE: L3 dendritic spine density differs by diagnosis and cortical region in SZ. Because dendritic spine size is associated with synaptic stability (large/medium spines) and plasticity (small spines), regional differences in the sizes of affected spines in SZ may reflect differing functional impairments in the primary sensory and association regions of the cortical visual-spatial working memory network.