Commentary — Cerebellar underdevelopment in the very preterm infant: Important and underestimated source of cognitive deficits

2.1 Cerebellar hemorrhage

Although the principal focus of this commentary is impaired overall development of the cerebellum, brief discussion of the major form of cerebellar parenchymal injury, i.e., cerebellar hemorrhage (CBH) is important, especially because the initial studies of CBH in the very preterm were the first to identify the relation of cerebellar affection to later cognitive-language-behavioral-socialization defects. Thus, in 2007 Limperopoulos and colleagues studied 35 premature infants (mean gestational age, 26 weeks) with isolated CBH [4]. Of this group, as predicted on the basis of the long-known association of the cerebellum with motor functions, 66% exhibited various neuromotor abnormalities consistent with cerebellar dysfunction. However, additionally, other deficits broadened the scope of apparent cerebellar-related sequelae, i.e., impaired receptive (37%) and expressive (42%) language, cognitive deficits (40%), socialization-behavioral deficits (34%), and abnormal autism screening (37%). Comparable data have been observed subsequently, especially most recently by Garfinkle et al. (n = 36) [10] and Boswinkel et al. (n = 218) [11]. Notably, the incidence of CBH in very preterm infants is approximately 16% [10], and although important, it appears likely that impaired cerebellar development as described next, is more pervasive, especially in infants of < 28 weeks’ gestational age.

2.2 Cerebellar underdevelopment

Impairment of growth of the cerebellum, in the absence of direct cerebellar parenchymal injury (e.g., CBH), is a prominent feature in preterm infants, especially those born very and extremely preterm. Identification of the abnormality is made most readily by quantitative measures with cranial ultrasonography (e.g., transcerebellar diameter) or more readily with MRI (e.g., transcerebellar diameter or regional and total volumes) [12]. The underdevelopment is associated with subsequent neurodevelopmental deficits, including cognitive-behavioral-language-socialization deficits (see later). Two broad mechanistic categories leading to the impaired development can be distinguished, i.e., direct effects and remote effects, as described next.

2.3 Cerebellar underdevelopment —direct effects

Direct effects on cerebellar development refer to those circumstances in which the effector acts directly on the cerebellum, most often at the level of the proliferating neuronal precursor cells of the EGL described earlier. The principal responsible effectors are blood products, glucocorticoid exposure, pain and opioid exposure, hypoxia-ischemia, systemic infection/inflammation, and preterm birth per se.

The likelihood that cerebellar underdevelopment can be related to blood products was shown initially by Messerschmidt et al. who described severe cerebellar growth failure in premature infants after severe IVH [13–15]. Hemosiderin deposition over the cerebellar hemispheres was shown by MRI. Hemosiderin is derived from blood by the following steps: hemolysis of red blood cells, formation of heme, conversion of heme to free iron (and biliverdin) by heme oxygenase, and formation of ferritin and then hemosiderin [16]. Free iron is toxic because it leads to the generation of reactive oxygen species, especially the hydroxyl radical by the Fenton reaction. The neuronal precursors of the EGL, underlying the hemolyzed red blood cells, are the target of the reactive oxygen species. Moreover, notably, cerebellar underdevelopment has been shown in association with low-grade IVH [17]. A careful study of 172 preterm newborns with serial MRI studies showed bilateral impairment of cerebellar growth at term-equivalent age associated with both mild and severe grades of IVH [18]. Because subarachnoid blood is a common finding in premature infants, even without overt IVH, this mechanism of iron facilitated free radical formation may be operative even in the absence of overt IVH [19].

Glucocorticoid exposure has been associated with impaired cerebellar growth in premature infants. Although antenatal glucocorticoid exposure is not associated with changes in cerebellar growth, post-natal exposure to betamethasone and dexamethasone has been followed by cerebellar underdevelopment [20, 21]. In one careful study of 224 very preterm infants, compared to 40 full-term infants, cerebellar volumes were smaller at term-equivalent age and at 7 years of age, with the largest deficits in the preterm infants of the earliest gestational ages [21]. The cellular site of the effect is almost certainly the granule precursor cells of the EGL, which are enriched in glucocorticoid receptors, activation of which leads to apoptosis of these crucial neuronal precursors [22]. Of additional importance in this context are the findings that (1) in premature infants, basal and peak serum cortisol responses in the first 2 weeks of life are highly variable, and (2) high serum cortisol levels documented in many infants likely represent continuing “stress” from respiratory and related disorders [23]. Taken together, the data suggest that the cerebellum of the very and extremely preterm infant may be exposed to high glucocorticoid levels from a variety of sources, exogenous and endogenous, and that these compounds may play a critical additive and/or central role in impaired proliferation of the EGL and thereby cerebellar underdevelopment.

Pain, stress and opioid exposure are associated with cerebellar underdevelopment in very and extremely preterm infants [24–26]. Studies have quantitated pain and stress in relation to the underdevelopment and have also identified morphine and fentanyl as negative effectors on cerebellar growth. That the site of the negative effects is the EGL is supported by experimental studies [27].

Hypoxia-ischemia and infection/inflammation have been associated with cerebellar underdevelopment in preterm infants [24, 28]. Experimental studies suggest that the EGL is the principal cellular target [29, 30]. In one such study glucocorticoids accentuated the deleterious effects of hypoxia [30]. However, because hypoxia-ischemia and infection/inflammation are so important in pathogenesis of cerebral lesions, especially cerebral white matter injury, in the very preterm infant, the role of remote (trans-synaptic) effects is difficult to separate from a direct cerebellar effect (see later).

Finally, premature birth and early extrauterine life may be directly deleterious to cerebellar development, in the absence of any appreciable cerebral injury and with varying degrees of control of other deleterious factors described earlier [21, 31–34]. Controlling for the variety of factors potentially deleterious to cerebellar growth is difficult. For example, in one experimental study in prematurely delivered baboons, ventilatory regimens seemed to play a role in cerebellar growth impairment [35]. Nevertheless, in one excellent experimental model (preterm piglets) preterm birth disrupted cerebellar development by impairing granule cell proliferation in the EGL [36]. The investigators concluded that preterm birth with precocious exposure to the ex-utero environment altered expression of key cerebellar developmental genes, affecting predominantly granule neuronal precursors in the EGL and Bergmann glia (along which the precursors migrate to populate the internal granule cell layer, as described earlier).

2.4 Cerebellar underdevelopment —remote effects

A second major mechanism involved in the cerebellar underdevelopment of the very and extremely preterm infant involves remote trans-synaptic affects, principally involving neuronal connections between the cerebrum and cerebellum. The major circuit involved begins in neurons of the cerebral cortex, axons of which traverse the vulnerable cerebral white matter, eventually synapsing on pontine nuclei in the brain stem. Pontine axons (so-called “mossy fibers”) then proceed (via the contralateral middle cerebral peduncles) to synapse on neurons of the internal granule cell layer of the cerebellum, the axons of which (“climbing fibers”) proceed to the dendrites of Purkinje cells. The latter cells send their axons to the cerebellar roof nuclei, principally the dentate, the axons of which proceed to the thalamus and then, via thalamocortical fibers, to multiple regions of the cerebral cortex. The tight relationship between cerebral cortical electrical activity and cerebellar growth was shown in a recent report of preterm infants from 30–40 weeks’ gestational age [37]. Lesions within this circuitry from the cerebral cortex to the cerebellum will lead to a loss of synaptic input and its associated trophic effects. The result is impaired development. Thus, it is not unexpected that there is a strong relation of cerebellar underdevelopment with such cerebral pathologies as cerebral white matter injury, periventricular hemorrhagic infarction and posthemorrhagic hydrocephalus [6, 38–42]. In one study in which the severity of white matter injury was evaluated relative to cerebellar volume deficit, a direct correlation was observed [40]. Consistent with the cerebellar circuitry just described and the trans-synaptic effects, infants with unilateral periventricular hemorrhagic infarction have been shown to exhibit diminished volume in the contralateral cerebellar hemisphere [38].

2.5 Cerebellar underdevelopment —neurodevelopmental outcome

Delineation of neurodevelopmental outcome attributable to cerebellar underdevelopment per se is hindered in those cases associated with prominent supratentorial disease, e.g. cerebral white matter injury, periventricular hemorrhagic infarction. In such cases, major motor deficits (spastic diplegia, hemiplegia, etc.) related to the cerebral lesions are prominent.

Of great interest is the relatively large number of cases of cerebellar underdevelopment, without major cerebral lesions, in whom neurodevelopmental outcomes relate principally to the cerebellar abnormality per se. Not unexpectedly, neuromotor deficits are apparent [21, 43–45]. However, most strikingly, disturbances of language development, cognition, executive and visual-spatial functions, mathematical computation, and IQ have been documented [21, 47]. The deficits have been observed as late as 7 and 10 years of age. The disturbances in cerebellar growth with diminished volumes have also been noted at similar later ages. These disturbances of higher neurological functions are consistent with affection of specific cerebellar circuits, involving cerebellar outflow via the dentate nucleus (see earlier) to the prefrontal cortex (executive functions), posterior parietal cortex (spatial cognition) and superior temporal cortex (language and complex auditory and visual processing). Disturbances of vermis are likely involved in the socialization defects observed in these children. Many of the features observed in infants with cerebellar underdevelopment are similar to the so-called cerebellar cognitive affective syndrome described initially in adults (see earlier).

It is reasonable to speculate that less severe disturbances of cerebellar development in extremely and very preterm infants contribute importantly to the spectrum of cognitive-behavioral-language-socialization deficits often attributed entirely to cerebral lesions. Herein lies a very fertile area for future clinical research.