President's Lecture - “A Foot in Two Camps”

I have called this presentation A Foot in Two Camps because for the greater part of my time in Neurosurgery and throughout the period that I have been a member of this Society for Research into Hydrocephalus and Spina Bifida I have been privileged and very fortunate to have had the opportunity to carry out laboratory research alongside my clinical practice. Inevitably, questions have arisen in one of the areas that have impacted on the other, and I would like to show how this has worked both clinically and in the laboratory in one of my special fields of interest, namely the growth and development of the brain.

It is often a single case that makes a striking impact clinically. Caroline, an 8-year-old girl was referred to the Neurosurgical Clinic at Booth Hall Children's Hospital in Manchester with a history of recent onset of headaches which were not particularly severe and did not even occur daily. She was one of 6 children; all her siblings had conspicuously good academic records at school and at university but Caroline had to struggle to keep up with her classmates although she had never been judged to be in need of extra help with her school work. In addition, her teachers had always thought she was rather slow and clumsy in the gym and was not much of a sportswoman on the playing fields. Clinical examination showed that she had a scar on her chest wall where surgery had been carried out when she was an infant to correct a Fallot's tetralogy; neurologically, there was nothing abnormal to find except for the suspicion of a right homonomous hemianopsia. A CT brain scan was carried out as part of the investigations of her headaches and it showed that the greater part of her left cerebral hemisphere was missing.

This scan made several questions spring to mind but the most interesting one was, could one date in her developmental history when this lesion occurred? And why had she so few neurological signs with so much of her left hemisphere missing? The explanation for why Caroline had so few neurological signs could have been due to a massive takeover of function of the missing parts of the brain by those that had developed normally, in other words by plasticity of the brain. That would account for her neurological status, but what about the timing of the onset of the development of the lesion? Was it an embryonic lesion, i.e. one occurring in the first eight weeks of pregnancy, or a fetal one arising any time during pregnancy after that time? In the literature, some researchers had reported that they had laboratory evidence indicating that the brain was capable of replacing lost neurones during early fetal life: an experimental fetal animal model with a teratogenically induced encephalocele and hydrocephalus at first sight appeared to show a histologically normal cortex and neurologically intact animals following intra uterine excision of the encephalocele and shunting of the hydrocephalus during the second trimester of pregnancy ([14]). These findings were inconsistent with well-established observations that human babies born prematurely during the second and third trimesters of gestation who had had a prolonged hypoxic and/or ischaemic episode or a parenchymal cerebral haemorrhage experienced loss of brain substance that was visible on imaging for the rest of their lives, and in addition, later in life, they frequently exhibited permanent neurological deficits which varied in severity but often included spasticity and hemiplegia. In these cases, it seemed that an insult to the brain in the second and third trimesters of pregnancy was followed by neither significant plasticity activity by the brain leading to normal or near normal neurological function nor to regeneration of lost brain tissue. But what did this reveal about Caroline's brain? It could be argued that whatever had happened to her had occurred even earlier in gestation than the episodes experienced by the premature babies; presumably, her incident had occurred at the very latest at the start of the second trimester, at a time when plasticity of the brain was presumably operating at a very efficient level in all areas except one, namely the visual cortex in the occipital lobe as evidenced by her persisting homonomous hemianopsia. But with this early insult was there any evidence that she had had some brain regeneration? It certainly did not look like it from her CT scans, however, there was no way of being sure simply by examining her or her scans. The only way to resolve the matter satisfactorily was to go into the laboratory and seek the answers there.

The first animal model my colleague at the time and I decided to investigate was one that involved making a traumatic lesion in the brain of a fetal rat. A needle stick injury was made in one of the cerebral hemispheres of fetal rats between days E17 and E21 which is a time when generation of neurones in the germinal matrix is occurring rapidly. The lesion was made by passing a hypodermic needle with an angled tip through the wall of the uterus and into one of the fetal cerebral hemispheres. This was done without disturbing the pregnancy in the majority of instances. After making the lesion, the pup was allowed to go to term and was delivered normally. The treated pups were allowed to survive for varying periods of time up to three months of age before being sacrificed. When their brains were examined it was found in all the long-term survivors that there was a cystic lesion at the site of the injury. Depending on the age of the animal at the time of sacrifice histological examination showed the presence of tissue necrosis, resolving haemorrhage, macrophages removing debris and scar tissue but no suggestion that neuronal regeneration had taken place in the area of the injury ([3]). It was therefore concluded that in rat fetus no brain regeneration occurred at the site of neuronal loss following trauma.

Because we wanted to specifically target the occipital lobe the method of making a lesion in the rat model was too crude and imprecise to allow this to be done. The size and site of the our lesions were very variable and could not be consistently reproduced, and therefore it was decided to extend our investigations to a fetal sheep model where a lesion could be made under direct vision ([4], [5]). In pregnant ewes of different gestational ages, under anaesthesia, the ewe's abdomen was opened and the head of one of the lambs was exposed through an opening in the uterine wall. A posterior craniotomy was performed on one side of the skull and the underlying occipital lobe was excised. The openings in the uterus and abdominal wall were repaired and the lamb was delivered spontaneously at term. The treated lambs appeared to be neurologically intact but it was impossible to examine their visual fields so it could not be determined whether they had a hemianopsia or not. All the lambs were sacrificed between the ages of 3 and 5 months. Examination of the brains showed that there was always a cyst in the occipital lobe at the site of the excision. Again examination of histological sections of the brain showed no evidence of regeneration of neurones. Moreover, there was evidence that profound changes had taken place in the rest of the operated hemisphere. The whole cerebral hemisphere on the side of the excision was significantly smaller than the opposite one. The white matter of the frontal lobe of the excised hemisphere was reduced in amount, but the thickness of the cortical mantle was often increased, suggesting that the development of the entire hemisphere had been disturbed by the excision of the fetus's occipital lobe. These changes were reflected on the external surface of the cerebral hemisphere by alteration of the normal gyral pattern in the hemisphere on the side of the excision. The ages of the fetuses at the time of excision ranged from the 63rd to 73rd day of gestation, the gestational period of the sheep being 145 days; which means that all the procedures were carried out in the second trimester of the sheep's pregnancy. The earlier in gestation the excision was carried out, the greater the changes seen in the operated hemisphere ([9]).

These results added confirmation to the view that neurones lost in the second and third trimesters of pregnancy were not replaced by regeneration. Although it proved impossible to assess whether the lambs had a homonomous hemianopsia, the fact that they certainly did not have other demonstrable abnormal neurological signs even though they had a cystic lesion in the brain and other changes made them resemble Caroline in many ways. In all of them a loss of brain tissue early in gestation had left a sizable cystic lesion, but plasticity of the brain and not regeneration of neurones had compensated for the loss of brain tissue in all areas except the visual cortex in the occipital lobe. These laboratory findings had helped to explain a lot about Caroline and, in particular, the signs demonstrable when she was examined neurologically.

Observations in the Fetal Management Unit at St. Mary's Hospital in Manchester have also led to questions that have made me go to the laboratory to seek answers. Starting in the early 1990s, a large number of fetuses with brain and spinal cord abnormalities were being referred to Fetal Management Units for ultrasonic evaluation. From the start, one of the commonest abnormalities seen was ventriculomegaly. In the early 1980s, it was suggested in some quarters that human fetuses with hydrocephalus would benefit from the drainage of cerebrospinal fluid from their ventricular systems on the grounds that this would allow their brains to develop more normally ([7]). A few units practicing fetal surgery in the United States and elsewhere had began inserting ventriculo-amniotic shunts into selected fetuses with ventriculomegaly. In total, about 40 or so were subjected to the procedure with varying results ([10], [11]). But why after this procedure did some of them do well and others not? By that time in the Fetal Management Unit at St. Mary's Hospital more than 40 fetuses referred with a diagnosis of ventriculomegaly had been seen and when they were followed up for even a short time it was obvious that some of them were doing well and others were not, in fact our findings looked uncommonly like the results that the fetal surgeons were getting.

Soon there were enough fetuses with ventriculomegaly to be able to start categorising them and to begin long-term follow-up studies on the natural outcome of the different groups ([8]). The first categories, the fetuses with ventriculomegaly were divided into, were complicated and isolated. Those in the complicated group were defined as having ventriculomegaly associated with a chromosomal or genetic abnormality, other brain and/or other organ malformations or evidence of infection, usually by one of the TORCH viruses. It soon became obvious from the follow-up studies that the outcome for this group was generally poor. Repeated ultrasound scans allowed the isolated ventriculomegalies to be further divided into the subcategories of increasing, stable and resolving groups. Follow-up showed that the outcome for the stable and resolving groups in the majority of cases was excellent whilst that for the progressive group was generally poor. Had these findings any implications for the outcome of the fetuses that had undergone intrauterine ventriculo-amniotic shunting? Could it be that those that did well had a stable ventricular size or one that would have returned to normal, irrespective of whether or not they had undergone shunting? Could it be assumed without doubt that those who had had a good outcome had it because of the beneficial effects of the procedure? The crucial question was, did all the shunted fetuses have true hydrocephalus? This could only be established if the same criteria were applied to them as are used to make the diagnosis post-natally, that is that hydrocephalus is an increase in the amount of cerebrospinal fluid (CSF) in the normal CSF pathways associated with an increase in the intracranial pressure which may not be elevated all the time although it usually is, and may not be increased by very much although, at least post-natally, it usually is. Hydrocephalus in the fetus, therefore, should only be diagnosed if ventriculomegaly is present together with the demonstration on repeated ultrasound scans that the ventricular atrial size is increasing progressively and there is an associated disproportionate increase in the head circumference.

The data collected in the St. Mary's Fetal Management Unit showed that true hydrocephalus was relatively uncommon compared to ventriculomegaly when hydrocephalus was assessed to be present using the criteria described above and if ventriculomegaly is defined as ventricular dilatation with a ventricular atrial measurement greater than 10 mm in a fetus whose head size is growing in proportion and isolated ventriculomegaly is defined as ventricular dilatation with a ventricular atrial measurement greater than 10 mm in a fetus whose head size is growing in proportion in whom there are no other abnormalities detected in the brain or the remaining organs of the body ([6], [8]).

Extended follow-up studies of the live-born fetuses up to the age of three years confirmed that the outcome for complicated ventriculomegaly, as indicated by earlier studies, continued to be uniformly poor. In the isolated group there was a marked difference in outcomes between those with progressive ventriculomegaly, the group into which the hydrocephalus cases fell, and those with stable and resolving ventriculomegaly, with only about 20 % of the progressive group having a good to fair outcome compared to the other two groups where nearly 80 % were doing well.

Looking specifically at the fetuses with hydrocephalus, one had to ask why they were uniformly doing so poorly. Postnatally, much of the damage caused by hydrocephalus is due to an increase in the intracranial pressure, but in the fetus, like in the neonate, the intracranial pressure never gets very high because the plates of the cranial vault are not fixed and come apart easily. Whilst intracranial pressure measurements cannot be made in the human fetus, it is probable that the same is true for them, after all the most likely reason why the ventricular atrial measurements in them are getting larger and the head size is increasing disproportionately is because there is an accumulation of CSF inside a ventricular system enclosed inside a readily distensible skull vault.

Hydrocephalus in the fetus, like in the post-natal paediatric population, is not a disease but a condition with many causes. In humans, even in a fetus with apparently isolated ventriculomegaly, when the ventricular atrial measurement is increasing and there is a disproportionate increase in the head size, the cause may be chromosomal, genetic, developmental or even due to acquired causes such as a hidden infection with one of the TORCH viruses. Another cause is blockage of the CSF pathways by a space-occupying lesion which not infrequently is an arachnoid cyst but one which is strategically placed and so small that it may be missed prenatally. Whilst undoubtedly the underlying cause of the hydrocephalus in some of these conditions could account for the failure of the cortical mantle to develop, there was a uniformity in the appearance of the hemispheres which suggested there might be a common underlying factor in all of them.

In the last 10 years in the laboratory of the Department of Biomolecular Science at UMIST there have been a number of postgraduate students and a postdoc looking at various aspects of cortical development in the fetal HTx rat. They confirmed the findings of others that cortical mantle development in the HTx rat fetus appeared to be normal up to the 18th day of gestation, but subsequently, after the development of aqueduct stenosis and the ensuing dilatation of the lateral ventricles, there was almost complete cessation of production of new neurones by the germinal matrix ([12]). This, together with dilatation of the ventricles, leads to the affected fetuses having a cortical mantle that is conspicuously thinner than that of unaffected animals. When stem and progenitor cells from the germinal matrix of affected Day 18 HTx fetuses were isolated and grown in culture medium alone or with added CSF from normal animals, it was found that the cells from the affected HTx fetuses were capable of multiplying supra-abundantly ([13]). On the other hand, cells from unaffected fetuses were inhibited and ceased to divide when CSF from affected fetuses was added to their culture medium. It was concluded from these results that it was a constituent or constituents of the CSF dammed up in the lateral ventricles that was inhibiting the division of the germinal matrix cells ([13]).

These findings offer an explanation for the appearance of the cortical mantle in the human fetuses with hydrocephalus assessed in the Fetal Management Unit. The onset of hydrocephalus in many cases is before the 18th week of gestation, that is at the very latest some time in the middle of the second trimester when it is known that the generation of neurones is occurring and should be continuing until about the 26th week. However, once hydrocephalus becomes established and CSF is dammed up in the ventricular system, inhibition by a substance or substances accumulating within the CSF stops neurone production, a situation that is common to all forms of hydrocephalus and is the most likely cause of thinning of the cortical mantle rather than the root cause of the hydrocephalus.

If it is accepted that cortical mantle development in these fetuses has been switched off by a substance that has accumulated in the CSF, could anything have been done to have removed the offending substance to allow the stem and progenitor cells to restart dividing and migrating? To date, no offending substances have been positively identified nor has it been determined whether removing CSF together with these substances from the fetal ventricular system would lead to the resumption of normal development of the cortical mantle. That needs to be established by future research.

What then of the other forms of ventriculomegaly that are seen in the Fetal Management Unit? Have the findings in the laboratory helped to explain what is happening to them? Do they offer an explanation why fetuses with stable and resolving ventriculomegaly generally have such a good outcome? First of all, it would seem unlikely that these fetuses have suffered a lengthy damming up of the CSF in their ventricular system because their cortical mantles are thick and there appears to be a normal gyral pattern on the cortical surface. Long-term follow-up studies suggest that the functional level achieved by these children later in life means that their brains are performing within the normal range ([6]). A possible explanation for their ventriculomegaly might be that the opening up of their ventricular system via the foramina in the fourth ventricle into the subarachnoid spaces is delayed with the result that there is a temporary accumulation of CSF within the lateral ventricles, but this is not accompanied by an accumulation of inhibitory material, or if it is, it is in insufficient amounts or not present for long enough to significantly affect the production of neurones by the germinal matrix. It is possible that the connection between the ventricular and subarachnoid spaces is established soon afterwards and even if neurone production is halted, it presumably is only temporary and restarts immediately after the CSF pathways open up.

These observations have important implications should it be suggested that intrauterine surgery might in the future have a role to play in the management of fetuses with hydrocephalus. First and foremost, before any procedure is contemplated, it is obviously imperative to establish that a fetus with large ventricles has hydrocephalus and not ventriculomegaly, since the latter does not require intervention and the majority of fetuses with ventriculomegaly do well without treatment. Intervention in them is not only unnecessary but exposes them to unnecessary risks including permanent brain damage. Laboratory studies would seem to suggest that if fetuses with hydrocephalus are to benefit from treatment, then intervention should be carried out as early as possible, preferably at the start of neurogenesis, since it is known that neurones lost later in gestation are not replaced, and there is no evidence to suggest that neurones that have not been produced at the right time can be replaced at a later date by an excess of production of neurones. Therefore it should not be assumed that removal from the ventricular system of dammed up CSF together with inhibitory agents comparatively late in gestation will lead to the production of a full compliment of neurones capable of migrating to their correct locations in the cortical mantle.

That hydrocephalus in the fetus has a profound effect on cortical mantle development is undoubted, but it should not be forgotten that hydrocephalus has other effects on the ventricular system as well. Postnatal stretching of the ventricular walls leads to intense changes in the ependymal lining of the lateral ventricles. These are striking when illustrated by use of the scanning electron microscope. Together with colleagues I carried out a scanning electron microscopic study of the ependymal lining of hydrocephalic neo-natal Hy3 Hy3 mice ([1], [2]). Normally, the luminal surface of the ependymal cells of the walls and roof of the lateral ventricle are covered in cilia. There are many theories about the function of the cilia. They are probably capable of moving CSF close to the ventricular wall, but whether they can move it from one chamber to another is disputed. Perhaps their function is to stir the CSF close to the ventricular surface to prevent accumulation of substances entering the lumen of the ventricle through the ependyma from the brain parenchyma. In hydrocephalic mice, stretching of the ependymal lining occurs as the ventricles dilate and this is accompanied by progressive loss of cilia. The ependyma of the lower parts of the wall of the lateral ventricle overlie the grey matter of the basal ganglia, the effects of stretching are minimal here even when gross hydrocephalus is present and only few cilia are lost, although in places this is enough to expose microvilli that are also present on the luminal surface of the ependymal cells. Higher up, the ventricular wall where the ependyma are overlying white matter, the effect of the stretching is much greater and many more cilia are lost. Eventually, even the microvilli begin to disappear. On the roof of the ventricle, where the effects of the stretching are greatest of all, the cilia and the microvilli are lost and even some of the ependymal cells are missing. What the effects of the missing cilia and microvilli are, is difficult to say.

Has having had A Foot in Two Camps been a useful experience? I believe it has. It has stimulated me to think about things in the laboratory and in a clinical setting and how each has interacted with the other in a positive manner. The signs are that in the future much more laboratory research is going be to be based on clinical problems. Grant funding bodies are more and more favouring research projects leading to a commercial outcome; so called “Blue Sky” research is getting less and less support; whether this is a good thing or not can be argued but at the end of the day the grant-giving bodies hold the money and they rule the day. The Society for Research into Hydrocephalus and Spina Bifida saw the value of mixed discipline groups getting together, exchanging ideas and imparting knowledge to one another almost from its inception, records of its annual meetings going way back show a healthy mix of clinicians and scientists talking to one another and informing each other of their findings in papers and posters. The Society should be proud that it predated government directives and grant-giving bodies' policies by many years by recognising the mutually beneficial values of science and medicine joining forces to solve problems. Long may this Society continue to encourage these two groups to work together. But to do so, both must use a common language so that each can understand the other and not be put off or be bored by the work being presented instead of being exciting and jolted into a thought-provoking state of mind. How much better it would be if after attending one of the meetings the clinicians were to say they now understood why something happens clinically after hearing one of the laboratory papers rather than saying they had understood nothing at all or the paper might just as well have been given in a foreign language. And how exciting it would be if the laboratory workers were to say after one of the clinical papers, “I wonder why that clinical situation always arises in that condition and now I must go into the labs and explore this further”; who knows where that might lead? Those things have been said in this Society in the past and will be said again in the future if we continue to listen to one another with open and enquiring minds.

I hope I have convinced you that whether you are primarily working in a clinical area or spend all your time researching in a laboratory, the papers you have heard over the years at the Society's meetings are not a boring waste of time, in each of them lies the potential for an answer to a burning question, a seed that could be implanted in your mind and lead on to a major advance in the management of patients, or the idea for a major research project. Cooperation lies not just in working together but in the exchange of ideas between open minds. This Society's proud history is that for nearly 50 years it has provided the opportunity for all of us to gather together annually and commune with one another for our mutual benefit and the development of our common interests. It is impossible to assess what the true worth of this Society has been over that time, but it is up to all of us to ensure that we make it possible that those that follow us will have a venue where they can achieve even bigger and better things in the next 50 years than were achieved in the last 50 by continuing the life and work of this remarkable Society.

References

• 1 Bannister C M, Mundy J E. Some scanning electron microscopic observations of the ependymal surface of the ventricles of hydrocephalic Hy3 and a human infant.  Acta Neurochirugica. 1979;  46 159-168
• 2 Bannister C M, Chapman S A. The ventricular ependyma - a scanning electron microscopic study of normal and hydrocephalic subjects.  Z Kinderchir. 1980;  32 371-372
• 3 Bannister C M, Chapman S A. Response of the fetal rat brain to trauma during the 17th to 21st days of gestation.  Dev Med Child Neurol. 1986;  28 600-609
• 4 Bannister C M, Chapman S A, Cranley J J, Turnbull I W. The outcome of unilateral occipital lobectomy in 12 foetal lambs in about the 100th day of gestation: a preliminary report.  Br J Neurosurg. 1987;  1 353-358
• 5 Bannister C M, Chapman S A, Cranley J J, Turnbull I W, Kulikowski J. Occipital lobectomies in foetal lambs.  Z Kinderchir. 1987;  42 (Suppl I) 5-9
• 6 Bannister C M, Russell S A, Rimmer S, Arora A. Pre-natal ventriculomegaly and hydrocephalus.  Neurological Research. 2000;  22 1-6
• 7 Birnholtz J O, Frigoletto F D. Antenatal treatment of hydrocephalus.  N Engl J Med. 1981;  304 1021-1023
• 8 Brown I M, Bannister C M, Rimmer S, Russell S A. The outcome of infants diagnosed prenatally as having cerebral ventriculomegaly.  J Mat Fetal Invest. 1995;  5 13-19
• 9 Chapman S A, Bannister C M. Effect of excision of occipital lobe tissue on about the 70th day of gestation on the growth and development of the sheep's brain.  Surg Neurol. 1989;  3 429-434
• 10 Clewell W H, Johnson M L, Meier P R, Newkirk J B, Hendee R W, Bowes W A, Hecht F, O'Kee D, Henry G P, Shikes R H. A surgical approach to the treatment of fetal hydrocephalus.  New England J Med. 1982;  306 1820-1825
• 11 Manning F A, Harrison M R, Rodeck and members of the International Fetal Medicine and Surgical Society C. Catheter shunts for fetal hydronephrosis and hydrocephalus.  New England J Med. 1986;  315 336-340
• 12 Mashayekhi F, Bannister C M, Miyan J. Failure of cell proliferation in the germinal matrix of the HTx rat.  Eur J Pediatr Surg. 2001;  11 (Suppl 1) 557-559
• 13 Mashayekhi F, Draper C E, Bannister C M, Pourghasem M, Owen-Lynch P J, Miyan J A. Deficient cortical development in the hydrocephalic Texas (H-Tx) rat: a role for CST.  Brain. 2002;  125 1859-1874
• 14 Michejda M, Petronas N, DiCharo G, Hogden C. Amelioration of fetal porencephaly in utero therapy of nonhuman primates.  JAMA. 1984;  251 2548-2558