Pathologically, white matter abnormalities correspond to areas of myelin thinning and gliosis and are often accompanied by lacunar (small holes) infarctions and small vessel atherosclerotic disease. White matter abnormalities relate most commonly to vascular disease and vascular risk factors. Although studies are incomplete, pathological correlation of leukoaraiosis and white matter abnormalities identified on neuroimaging would be expected to help determine the pathogenesis of leukoaraiosis and white matter abnormalities. One problem is that fewer white matter abnormalities are visible postmortem when compared with pre-mortem MRI (Fazekas et al 1991).
The presence of periventricular region leukoaraiosis has been correlated with decreased myelin content (Sze et al 1986; Leifer et al 1990; Grafton et al 1991; van Swieten et al 1991; Chimowitz et al 1992; Fazekas et al 1993; Scarpelli et al 1994; Moody et al 1995), loss of ependymal cell layer and reactive gliosis at the tip of the frontal horns, (Sze et al 1986; Jungreis et al 1988; Chimowitz et al 1992; Fazekas et al 1993; Scarpelli et al 1994), increased periependymal extracellular fluid content, axonal loss, and presence of atrophic axons (Sze et al 1986). Some authors have also found enlarged perivascular spaces at this periventricular location (Grafton et al 1991). However, small periventricular lesions exist in all age groups (including newborns) (Sze et al 1986; Moody et al 1995); therefore, smaller lesions may not represent a true abnormality.
Pathological studies of white matter abnormalities remain scanty. They have suggested a relationship to frontal atrophy, ventriculomegaly, and presence of reactive astrocytes in the frontal periventricular white matter, as well as increased arteriolar wall thickness (Whitman et al 1999). An autopsy study examined brain tissues from pathologically confirmed Alzheimer disease with tissue processed for ApoE genotyping. The presence of myelin loss in these brains was not related to the ApoE4 genotype, nor was it related to the presence of pathologically confirmed arteriosclerosis. (Tian et al 2004).
The histological correlates of deep subcortical white matter abnormalities are even less consistent than for periventricular leukoaraiosis. Punctate abnormalities correspond to enlarged perivascular spaces (Chimowitz et al 1992; Scarpelli et al 1994), lacunes (Braffman et al 1988b; Marshall et al 1988; Munoz et al 1993), demyelinating lesions, brain cysts, and congenital diverticula of the lateral ventricles (Braffman et al 1988b). More diffuse lesions within the centrum semiovale have been compared to myelin rarefaction sparing of the U fibers (Revesz et al 1989; Chimowitz et al 1992), sometimes accompanied by reactive astrocytes (Fazekas et al 1993), as well as diffuse vacuolization of the white matter (Munoz et al 1993). The myelin rarefaction that occurs is not a true demyelination, as the process also will involve axonal destruction (Awad et al 1986; Lotz et al 1986).
Subcortical white matter abnormalities disrupt short corticocortical fibers, whereas periventricular leukoaraiosis may lead to damage within regions containing closely packed long association fibers connecting distant cortical areas (de Groot et al 2000). Using the hypothesis that white matter abnormalities lead to a disconnection syndrome (Mesulam 1990; Wolfe et al 1990), white matter abnormalities are more likely to disrupt local neuronal networks, whereas periventricular leukoaraiosis is more likely to impair cognitive functions, requiring the coordination of multiple distinct cortical areas.
There is anatomical, histopathological, clinical, and experimental evidence that at least some white matter abnormalities are ischemic in origin (de Reuck 1971; Fazekas et al 1987; Fazekas et al 1993; Erkinjuntti et al 1996; Longstreth et al 1996; Pantoni and Garcia 1997). Experimental animal studies have reproduced some aspects of human white matter lesions (Pantoni and Garcia 1997). Some studies have implicated leaks in the blood-brain barrier as at least partially pathogenic for the appearance of white matter abnormalities (Kemper et al 2001). Alterations in cerebrospinal fluid circulation may play a partial role in the development of white matter abnormalities (Kimura et al 1992).
Regardless of the location of white matter abnormalities, deficits in frontal lobe metabolism detected by PET are found in predemented diabetic patients (Whitman et al 2001), perhaps related to structural or metabolic damage to subcortical axons. Abnormalities in MRI T2 values over brain regions with white matter abnormalities could suggest changes in water content or a modification of quantity of molecules or proteins that would produce abnormal and T2 signal. Although early diabetes is associated with mild loss of water content in the brain (Tullberg et al 2004), chronic hyperglycemia, as well as diabetic ketoacidosis, is associated with brain water content preservation or increase, possibly due to osmoprotective molecule production (Tornheim 1981).
Neuropathologic studies of the brains of nonhypertensive elderly patients have demonstrated prominent frontal atrophy and ventriculomegaly. In addition, markedly reactive astrocytes were found in periventricular white matter of most patients with white matter abnormalities and gait imbalance. An increased arteriolar wall thickness was demonstrated in some patients with white matter abnormalities as well (Whitman et al 1999).
Although myoinositol concentrations are significantly increased in the frontal white matter of patients with diabetes as compared to healthy controls, its importance and possible relationship to diabetes-associated gait apraxia and cognitive impairment remains uncertain (Ajilore et al 2007).
An animal model for longstanding diabetes has been associated with the presence of cerebral atrophy and WMA (Toth et al 2006). Both MRI volumetric assessments and brain weight revealed brain atrophy in long-term diabetic mice as compared to littermate controls. Furthermore, leukoencephalopathy with evidence of MRI hyperintensities over the hippocampus, thalamus, putamen, corpus callosum, and internal capsule were associated with pathologically illustrated myelin loss or pallor. These pathological changes were also associated with time-related development of cognitive changes during behavioural testing. A possible pathological marker for these changes is the increased expression of RAGE (the receptor for advanced glycation end products), which was found to be increased dramatically within sites of white matter damage. RAGE expression was elevated within neurons, oligodendrocytes, astrocytes and microglia. Meanwhile, RAGE null diabetic mice demonstrated significantly less neurodegeneration when compared to wild-type diabetic mice (Toth et al 2006). Further studies examining potential signaling pathways of RAGE are underway, as well as studies to examine blockade of the RAGE pathway with the competitive decoy, soluble RAGE (sRAGE).