For the first time the current study has carried out a longitudinal, prospective assessment of the plasma levels of MMP-9 in a homogenous cohort of congenitally deaf implanted children, and specifically excluding a number of factors that might potentially affect cochlear implantation outcomes.
We have shown that an MMP-9 plasma level of 150 ng/ml, measured at cochlear implantation, appears to be a reasonably effective threshold for differentiating good and poor performers later in their auditory development. We also find that the predictive value of MMP-9, measured at implantation, on the LEAQ score at 18 months also holds for the plasma level of MMP-9 measured at 8 months. However, LEAQ scores measured at cochlear implantation and at later intervals did not correlate with corresponding measurements of MMP-9.
Grouping of good and poor performers
Based on MMP-9 levels measured at cochlear implantation we have isolated two clusters of performers. Cluster 1 were those children who had an MMP-9 plasma level of below 150 ng/ml just before they started speech and language rehabilitation with their CI, and after 18 months they were among the best performers. Cluster 2 were those children who had higher levels of MMP-9 at cochlear implantation, and here one can see a considerable variance in their LEAQ scores at the end-point of observation. Because in both clusters the mean values of age at CI activation carry a variance, one cannot attribute a direct effect of age on the variance of LEAQ score at 18 months. These findings lead us to conclude that an MMP-9 plasma level below 150 ng/ml at the time of cochlear implantation predisposes a child who has been diagnosed with DFNB1-related deafness to be a good performer, and is highly likely to reach full auditory development 18 months later. The same pattern exists for MMP-9_8 and LEAQ_18.
Nevertheless, there was one child in cluster 1 who started with an MMP-9 serum level of 188 ng/ml and who went on to score the maximum number of points in LEAQ 18 months later. Detailed analysis of this case revealed that the child was raised by very committed parents and was introduced into a rich, regular rehabilitation program. On LEAQ, the child had already scored 33 points at the 8 month follow-up, which suggests that the parents’ attitude played a big part in the treatment outcome.
Temporal dynamics of MMP-9 in plasma
An advantage of a longitudinal study is that it allows one to analyze the impact of certain features, and how they interact over time. In our material, prospective observation of protein plasma levels in both clusters enabled us to see that MMP-9 concentrations decreased over time, although there were no significant differences between them. The natural changes in concentrations of these proteins in plasma and brain tissue of human neonates and toddlers remain unknown, either under healthy or pathological conditions, and for this reason we must rely on data from animals and human adults. Results of animal studies indicate that the protein is highly abundant, especially in the juvenile brain under dynamic development (11, 14). In rats, MMP-9 concentrations in the brain, particularly in the hippocampus, are highest in the earliest postnatal period and decrease into adulthood (38, 39). Kudo et al. reported on MM-9 protein plasma levels in human adults, and their results from healthy volunteers are much lower than in our subjects, showing mean values of 26.4 ng/ml versus 121.4 ng/ml here (40).
Brain–blood plasma relation
There are several questions arising from our results. Firstly, the peripheral levels of MMP-9 might not relate to the protein’s level in brain neuropil at all. However, some evidence emerging from studies of mice shows that MMP-9 cleaves collagen IV – a major component of the basal membrane of the cerebral vascular endothelium – and may thus change the permeability of the brain–blood barrier (BBB) (41). It has been shown that levels of MMP-9 in animal brain cortex, cerebrospinal fluid, and serum are simultaneously coupled (38, 39, 41, 42). In this context, despite existing evidence that the plasma levels of the protein can be influenced by its secretion from endothelium, we assume similar dynamics between brain and blood in humans (41, 43). For now, no research tool is available to assess MMP-9 protein concentrations in the cortex of the human infant brain.
We suggest, although with extreme caution, that the decrease of MMP-9 levels observed in our clusters might be interpreted in analogy with data on Fragile X syndrome (FXS), a condition associated with cognitive deficits, anxiety, mood imbalance, and increased MMP-9 plasma levels (27). A clinical trial of minocycline, which lowers MMP-9 levels, has shown that the drug can attenuate clinical features of FXS (28). It is therefore possible to conclude that large decreases in peripheral MMP-9 levels, either pharmacologically induced or naturally occurring, might correspond with better neurocognitive conditions.
Variance of MMP-9 plasma levels
Another question that needs to be addressed here is the observed variance across our cohort of MMP-9 plasma levels measured at cochlear implantation. It is known that MMP-9 activity remains under tight and complex regulation, both temporally and spatially, largely through gene expression, mRNA, and protein translation (14). MMP-9 is a pleiotropic enzyme involved in numerous processes with a long list of targets, including growth factors, cell surface receptors, and cell adhesion molecules, which we think may explain the variance (11, 12, 42). Functional gene polymorphisms also affect gene expression patterns, and, depending on the combination, protein levels may be different (14, 26, 42).
Limitations
It is important to underline that no data currently exists on the relation between MMP-9 concentration in brain and blood. However, our aim was not to elucidate the exact mechanisms of MMP-9 involvement in synaptic plasticity, instead we were looking for a potential biomarker of neuroplasticity. At present, there are no objective tests for assessing auditory development in children up to 2 years old. We therefore have to accept the parents’ subjective rating as incorporated into the LEAQ score. We have not taken into consideration the impact of family support on the likelihood of successful rehabilitation (2). Detailed observation of rehabilitation would take much time and would require more frequent follow-ups. Ethnically, our cohort comprised only a Caucasian (Polish) population, so more diverse studies would be of great value in confirming the broader character of our results.