EXPERIMENTAL NEUROLCSY 84,207-2 18 ( 1984)

Neuronotrophic Activities in Cerebrospinal Fluid of Head Trauma Patients

FRANK M.LONGO,IVAN SELAK, JOHNZOVICKIAN, MARSTON MANTHORPE, SILVIO VARON, AND HOI SANG U’

Departments of Biology and Surgery, Division of Neurosurgery, School of Medicine, University of CaliJonia, San Diego, La Jolla. California 92093

Received August 29. 1983; revision received November II, 1983

Neuronotrophic factors (NTFs) are agents required for neurons to survive in tissue culture. In this study, we investigated the presence of NTFs in cerebrospinal fluid (CSF) of patients with central nervous system (CNS) injury. Cerebrospinal fluid was collected from I5 patients with acute CNS lesions in whom ventricular catheters had been placed to monitor and to facilitate the control of intracranial pressure. Neurono- trophic activity within the CSF was assayed using cultures of neurons derived from fetal rat hippocampus and embryonic chick cerebral cortex. Cerebrospinal fluid from all IS patients contained NTFs which supported the survival of rat hippocampal neurons. Survival of chick cortex neurons was supported by eight of nine CSF samples. In the 11 patients from whom consecutive CSF samples were available, NTF activity assayed in rat hippocampal cultures tended to decrease during the first several days after CNS injury. In CSF collected from three patients by lumbar puncture for diagnosis of “nontraumatic” conditions, no NTFs were detectable. NTFs supporting hippocampal neurons were also detected in extracts of blood clot obtained from normal volunteers. Neuronotrophic activity in the CSF was heat sensitive, nondialyzable, and macro- molecular, suggesting its association with a protein(s). These observations suggested that (i) NTFs are detectable in human CSF after CNS injury, (ii) NTFs appear in response to the injury itself, and (iii) at least some human NTFs can support the survival in culture of nonhuman CNS neurons.

Abbreviations: CNS-central nervous system, CSF-cerebrospinal &rid, NTFs-neurono- trophic factors, NGF-nerve growth factor.

’ I. S. is Aspirant du FNRS (Belgium), University of Liege, Liege, Belgium, Department of Neurology, School of Medicine, which is his present address. This work was supported in part by a grant from the National Institutes of Health (16349) to Silvio Varon. The authors also thank Diana Rosati for her excellent secretarial assistance. Reprint requests should be sent to Dr. U, Division of Neurosurgery, UC San Diego Medical Center, 225 Dickinson St., San Diego, CA 92 109.

00144886184 $3.00 Copyright 8 1984 by Academic Rcss, Inc. All rights of reproduction in any form rewwd.

208 LONG0 ET AL.

INTRODUCTION

Neuronotrophic factors (NTFs) are macromolecules required for the sur- vival and neuritic regeneration of axotomized neurons in most neuronal culture systems (16). The physiologic function of NTFs in vivo, especially with respect to neuronal regeneration, is only beginning to be elucidated. Exogenously supplied NTFs, such as nerve growth factor (NGF) (2, 5, 14, 19) and other trophic agents (13) enhance neuronal regeneration after CNS injury in experimental animals. The inhibition of adrenergic reinnervation in the peripheral nervous system (PNS) by the administration of antiserum to NGF further suggests that endogenous nerve growth factor supports neu- ronal regeneration in vivo (2, 4).

Recent investigations demonstrated that NTFs supporting the survival of cultured peripheral neurons accumulate in wound cavities made in the en- torhinal-occipital cortex of developing rats (10, 11) and in various brain regions of adult rats (10). In those studies, neuronotrophic activity was very low in noninjured brain tissue and in the wound cavity 1 day postlesion. This activity, however, increased 15 to 300-fold during the subsequent 15 days. At the same time, survival of transplanted fragments of rat embryo septum or corpus striatum was poor when placed in such wound cavities immediately after the cavity was made. When implantation was delayed, survival and regeneration of cholinergic neurons within the implanted fmg- ments increased. This corresponded to the increase in NTFs present in the cavity. These observations suggest that NTFs detected using in vitro methods may also influence neuronal survival in vivo; however, a direct cause and effect relationship remains to be established.

In light of the potential effect of NTFs on neuronal survival and regeneration in vivo after injury of nervous tissue, we investigated the presence of NTFs in human cerebrospinal fluid (CSF) that follows CNS injury. The following points were assessed: (i) the existence of CNS neuron-supporting NTFs in the CSF of patients with or without CNS trauma, (ii) the temporal relationship of such NTFs to the traumatic event, (iii) the molecular nature of trauma- related NTFs, (iv) the occurrence of NTFs in normal human blood clots, and (v) the species specificity of such NTFs. Samples of CSF and blood clot extract were assessed with a recently developed NTF assay using CNS neurons derived from embryonic rat hippocampus and embryonic chick cerebral cortex (3).

MATERIALS AND METHODS

Collection of Cerebrospinal Fluid From Patients with Injury of the Central Nervous System. The study included 15 patients admitted to the Neurosurgical Service at the University Hospital, University of California, San Diego, during

NEURONOTROPHIC ACTIVITY IN TRAUMA CSF 209

a l-year period (Table 1). Fourteen of the patients were admitted for closed head trauma and received ventricular catheters within 1 day after admission to facilitate the monitoring and control of elevated intracranial pressure. One

TABLE 1

Characteristics of 15 Patients with Head Trauma

Patient Etiology of CNS injury

Glasgow coma score on admission (15) CT scan findings

SC MVA (motor vehicle accident)

RW MVA

MH

BH

MVA

MVA

DS

EK

MVA

MVA

GB

AH

MVA

Rodeo accident, depressed cranial fracture

TN

RL

MVA

Fall

KW

WN

MVA

EJ

MG

RD

Ruptured aneurysm and repair

Trauma (MVA?)

Bicycle fall and wary

MVA

8

8

1

6

Small intraparencymal hemorrhage

Subdural hematoma, subarachnoid blood, cerebral contusion

Brain stem hemorrhage

Frontal contusion, subdural blood

Cerebral contusion

Bifrontal contusions, subarachnoid blood

Within normal limits

Epidural hematoma, subarachnoid hemorrhage, intracerebral hemorrhage

Within normal limits

Massive intraparenchymal and intraventricular hemorrhage

Cerebral contusions

Subarachnoid blood, intracerebral blood, clipped aneurysm

IntracerebraI hemorrhage

Cerebral contusion

No intracranial hemorrhage

210 LONG0 ET AL.

patient had a ventricular catheter placed immediately after surgery for an intracranial aneurysm. The CSF was collected from ventricular catheters by allowing about 100 ~1 fluid to drain into a sterile test tube. Samples were stored at 4°C from 1 to 10 days. Aliquots were made and stored at -76°C for several days to several weeks prior to the trophic factor assay.

Collection of “Nontrauma” Cerebrospinal Fluid. The CSF which had been collected by lumbar puncture for diagnostic studies was stored at 4°C for 1 to 10 days after completion of the necessary clinical tests. Aliquots were then made and stored at -76°C for several days. Samples were obtained from three patients with spinal myeloma, suspected multiple sclerosis, or dementia, respectively.

Blood Clot Extracts. To determine if trauma-associated CSF neuronotrophic factors could originate from bleeding within the CNS, the presence of NTFs in blood clot was assessed. Five milliliters blood were drawn by antecubital venopuncture from each of five normal volunteers. Samples were stored 3 h at 4°C after clotting. Extracts of the clots were prepared in phosphate- buffered saline (10% clot: 90% saline, v/v) by homogenization in a glass homogenizer. The resulting 10% blood clot extracts were stored at -76°C for several days before assaying.

Cell Cultures for Neuronotrophic Factor Assay. The culture medium con- sisted of Eagle’s Basal Medium (EBM, Gibco) supplemented with 26.4 mA4 NaHC03, 33.3 m&I ~-glucose, 2 mA4 L-glutamine, and 100 U/ml penicillin. The Nl supplement (6) consisted of 5 &ml insulin, 5 &ml transferrin, 10m4 M putrescine, 2 X low8 A4 progesterone, and 3 X lo-* A4 selenium (as Na2SeOs).

Neurons derived from rat hippocampus and chick cerebral cortex were prepared and cultured as described elsewhere (3). Briefly, the septum from 18-day-old rat fetuses and the cerebral cortex from day 8 chick embryos were dissected and placed in Hank’s balanced salt solution. Each tissue was in- cubated 10 min at 37°C in calcium- and magnesium-free saline (Hank’s salts) and then 20 min at 37°C in 0.08% trypsin in calcium- and magnesium- free saline. The tissues were washed twice in culture medium containing 1% ovalbumin (Sigma) and dissociated in 2 ml of the same medium by 20 passages through the flamed tip of a siliconized glass Pasteur pipet. The cell suspensions were then filtered through Nitex (20 pm) and diluted with culture medium containing the Nl supplement,

Tissue culture wells (6 mm diameter, Costar 3596) were precoated 3 h at room temperature with 50 ~1 polyomithine (0.1 mg/ml in 15 mA4 borate buffer, pH 8.4, Sigma), rinsed twice with 50 ~1 HzO, and supplemented with 50 ~1 serum-free conditioned medium from RN22 Schwannoma cell cultures which contained a polyomithine-binding neurite promoting factor (1). The presence of neurites promoted by this substratum factor facilitated the rec-

NEURONOTROPHIC ACTIVITY IN TRAUMA CSF 211

ognition of neurons. After an overnight incubation at 37’C in 5% CO*-95% air, the medium was removed and the coated wells were rinsed twice with 50 ~1 culture medium. The washing medium was then replaced by 50 ~1 culture medium containing serial two-fold dilutions (beginning at 1:4) of CSF or 10% blood extract samples and the Nl supplement. Fifty microliters of the neuronal cell suspension (2000 cells/well) was added to this medium. The cultures were incubated at 37°C in a humidified 5% CO*-95% air at- mosphere. After 24 h, each culture was fixed 30 min at 22’C by the addition of 200 ~12% glutaraldehyde in culture medium before the number of surviving neurons was determined. For each well, neurons were counted in two-diameter strips (0.5 X 6 mm) each representing 10% of the total surface area. Under these culture conditions with a low initial seeding density, no CNS neuron survived on its own for 24 h or longer without the addition of exogenous trophic or metabolically essential factors (3). This system therefore provided the appropriate means for the recognition and analysis of exogenous survival- promoting (i.e., trophic) agents for these neurons.

Quantijication of Neuronotrophic Activity. Cells in our cultures were clas- sified as neurons under phase-contrast microscopy by morphologic criteria described elsewhere (3). These cells corresponded to cells identified as neurons by indirect immunofluorescent techniques using tetanus toxin. In a dual counting procedure, more than 90% of the cells in culture demonstrated a “neuron-like” morphology, and of these, more than 90% stained positively with the tetanus toxin reaction. These results therefore support the contention that the cultures consisted mostly of neurons which could be identified with confidence using morphologic criteria alone.

The number of surviving neurons in each well was determined and plotted against the dilution of CSF or extract sample. The dilution supporting one- half the maximum neuronal survival was defined as the number of trophic units per milliliter (TU/ml) of the original fluid or extract sample. Without the addition of neuronotrophic factors in these assays, no neurons would survive for even 24 h. All samples from a given patient were assayed with a single-cell suspension in order (i) to minimize variation due to the use of different cell suspensions, and (ii) to derive the most meaningful comparison between data from different time points because the amount of CSF obtainable from each time point would permit the performance of only one assay. The ability of each neuronal dissociate to respond to NTFs was confirmed by adding medium conditioned by astroglia [shown elsewhere to contain trophic agents for CNS neurons (3)] to control wells.

Molecular Characterization of Neuronotrophic Activity (Table 2). Five CSF samples (from four patients) with the highest NTF titers were selected for molecular characterization. The first aliquot of each sample was used as the control. A second aliquot was heated 10 min at 90°C to determine heat

212 LONG0 ET AL.

stability. A third aliquot was dialyzed 30 min against 200 ml phosphate- buffered saline in a dialysis membrane (molecular weight cutoff at 12,000 Da) at room temperature. A fourth aliquot was filtered through a PM10 ultrafiltration membrane. Both the retentate and eluate were assayed for NTFs. These last two tests were aimed at estimating the molecular weight of the active constituents and, thus, distinguishing between macromolecular trophic “factors” and low-molecular-weight trophic agents such as essential metabolites ( 17).

RESULTS

The CSF from all 15 patients with traumatic CNS lesions contained neu- ronotrophic activity supporting the survival of rat hippocampal neurons (Fig. 1A). In cultures containing trauma-associated CSF, a maximum of 60 to 70% of the seeded cells (which included neurons as well as other cells) survived and extended neurites. In the absence of neuronotrophic supplements, no surviving neurons were observed after 24 h in culture. The CSF from eight of nine patients also contained NTFs supporting chick cerebral cortex neurons (Fig. 1B).

In this study, consecutive CSF samples were obtained from 11 of 15 patients. Hippocampal neuronotrophic activities of CSF samples collected at various times postinjury are diagrammed in Fig. 2. The temporal patterns of NTF titers for individual patients varied markedly although the general trend was a decline of activity. In each patient, the first CSF sample collected within 1 day of the CNS injury contained considerable NTF activity which was consistently higher than the last sample measured. The CSF neuronotrophic activity tended to decrease during the initial 1 to 2 weeks after injury; however, two patients had a marked but transient increase in activity above the initial level (KW and RW). These transient increases were not accompanied by any apparent change in the clinical status such as reinitiation of an intracranial bleed or acute brain swelling. In the remaining four trauma patients, only single CSF samples were obtained. All four samples were collected within 1 day after trauma and contained neuronotrophic activity supporting rat hip- pocampus neurons. The NTF activity, in neuronotrophic units per milliliter CSF, were: BH, 75; RL, 460; EJ, 45; and MG, 240. In an assay of hippocampal neuronotrophic activity, no change was observed in CSF specimens stored for long durations (weeks) at 4°C.

The CSF collected from nine of the trauma patients described in Table 1 was also assayed for NTF with chick cerebral cortex neuron cultures. Con- secutive samples were available from seven of those patients, and six of them contained chick cerebral cortex NTFs which decreased after trauma (Fig. 3).

FIG. I. Cultures of rat hippocampus and chick cerebral cortex neurons. A-rat hippocampal neurons cultured in the presence of cerebrospinal tluid (diluted 1:4) collected after CNS trauma. Neurons survive and extend neurites. B--chick cerebral cortex neurons cultured in the presence of cerebrospinal fluid (1:4) collected after CNS trauma. Note similar neurite extension. Calibration bar = 50 pm.

213

214 LONG0 ET AL.

DAYS POST-TRAUMA

FIG. 2. Quantitation of neuronotrophic activity in cerebrospinal fluid collected from I 1 patients (identified by initials) at various times after CNS injury. Serial dilutions of CSF were assayed with rat hippocampus neurons. The fluid dilution supporting one-half the maximum neuronal survival was defined as the number of trophic units per milliliter of fluid (TU/ml).

This was similar to the observation with the hippocampal neuronal assay. Samples from the last of those seven patients, patient GB, had no detectable chick cerebral cortex NTFs but these samples did contain hippocampus NTFs (Fig. 2). The CSF from the two patients from whom only single samples were obtained both contained chick cerebral cortex NTFs (MH, 200 and BH 3 TU/ml CSF). The quantity of neuronotrophic titers detected with chick cerebral cortex neurons was generally less than that detected by rat hippo- campus neurons, 0 to 200 Tu/ml versus 0 to 700 TU/ml.

Neuronotrophic activity for hippocampal neurons was assayed in aliquots of CSF after heat treatment, dialysis, and filtration (Table 2). The NTF titers were significantly reduced by heat treatment (P c 0.05, Wilcoxon signed

NEURONOTROPHIC ACI-1Wl-Y IN TRAUMA CSF 215

I2 34 5 6 7 8 9 IO II 12

DAYS POST- TRAUMA

FIG. 3. Quantitation of neuronotrophic activity in cerebrospinal fluid collected from six patients (identified by initials) at various times after CNS injury. Serial dilutions of CSF were assayed with chick cerebral cortex neurons. The fluid.dilution supporting one-halfthe maximum neuronal survival was defined as the number of trophic units per milliliter of fluid (TU/ml).

rank test for matched pairs, N = 5), but not significantly changed by dialysis (P < 0.05, Wilcoxon signed rank test for matched pairs, N = 5). All NTF activity recovered after ultrafiltration remained in the retentate indicating the association of NTF activity with a species of molecular weight greater than approximately lo4 Da. This suggested that the hippocampus NTF is a protein.

No hippocampal neuronotrophic activity was detected in any of the three “nontrauma” CSF samples. Extracts of blood clot, however, contained con- siderable NTF titers supporting hippocampus neurons. Ten-percent extracts contained 1300 TU/ml to 4200 TU/ml with an average of 2260 TU/ml. Trophic activities present in blood components are currently under inves- tigation in this laboratory (18).

216 LONG0 ET AL.

TABLE 2

Neuronotrophic Activity of Cerebrospinal Fluid after Various Treatments

Treatment NTF titer after treatment” (%)

None (control) 100 Heath 0 Dialysis’ 85 f 26 Filtrated 0 Retentate 95 + 21

LI In this analysis, the NTF titer of the untreated CSF sample (i.e., the control) is taken as 100%. The activity retained by replicate aliquots after treatment indicated is a percentage of the control activity.

b Heat treatment was 10 min at 90°C. c Dialysis was 90 min at room temperature against saline, through a 12,000-Da cut-off membrane. d Filtrate and retentate were fractions collected by ultratiltration on a PM 10 membrane.

DISCUSSION

Neuronotrophic factors have been studied mainly with respect to peripheral neurons (PNS) (7, 9-11, 16). Cultures of intrinsic CNS neurons were only recently developed in a manner suitable for quantitative analysis of exogenous trophic agents (3, 17). Thus, some caution must be exercised in comparing trophic macromolecules and their behavior in CNS and PNS contexts.

In this study, we demonstrated that macromolecular NTFs directed to CNS neurons are found in human CSF in association with CNS trauma. The demonstration of a significant activity soon after acute CNS trauma and the tendency for those NTFs to decline in the ensuing days following trauma together with their absence in “nontrauma” CSF suggest that they appear as a response to the acute traumatic event itself. The gross temporal pattern of NTF activities in the CSF is similar to observations concerning PNSdirected NTFs that follow peripheral nerve injury, where peak activities are also attained immediately after the acute event (9). The NTFs in the CSF may be synthesized in response to injury or they may already be present and are released upon injury. There are several potential sources of those NTFs. The CNS cells may synthesize NTFs. Culture media conditioned by glia contain macro- molecular NTFs for PNS neurons and low-molecular-weight agents for CNS neurons (7,8, 17). The latter are also present in media conditioned by neurons (3). Alternatively, trauma-induced NTFs in the CSF may originate from blood hemorrhaging into the CSF as a result of head trauma. This is suggested by the relatively high NTF titers detected in blood (18).

NEURONOTROPHIC ACTIVITY IN TRAUMA CSF 217

The NTFs supporting peripheral neurons were shown to accumulate in rat cortex lesions; however, in contrast to the NTFs in human CSF, their titers increased during the first 1 to 2 weeks after cortical injury (11, 12). One possibility accounting for the different temporal patterns of human CSF and rat cortex NTFs is that NTFs from human CSF assayed in this study were directed to intrinsic CNS neurons, whereas, rat cortex NTFs were directed to PNS and spinal cord (presumptive motor) neurons. Another possible reason for this temporal difference may reside in the local conditions of the different sites from which the NTF-containing material was collected. As described in Table 1, many of the human brain lesions were situated at some distance from the ventricles where CSF was collected. In the ventricular system, there is a continuous turnover of CSF. Substances synthesized or released as a response to acute trauma may thus be diluted in time with fresh CSF. This would account for the declining NTF activities observed. In contrast, NTFs in rat cortical lesions were derived from extracts of Gelfoam sponges implanted directly into the wound or from adjacent brain tissue ( 11,12). Those conditions would favor prolonged local accumulation of NTFs in and near the injured tissue.

The discovery of NTFs in human CSF in association with brain injury has several important implications especially in the regeneration of nervous tissue in vivo and in the survival and regeneration of brain tissue grafts. These observations will assume added importance should future studies demonstrate a positive correlation between NTF activity and clinical status. Association between increased NTF activity and clinical improvement would have im- mense therapeutic significance.

REFERENCES

1. ADLER, R., M. MANTHORPE, S. SKAPER, AND S. VARON. 198 1. Polyomithine-attached neurite-promoting factors (PNPFs): culture sources and responsive neurons. Brain Res. 206: 129-144.

2. ALOE, L., R. CALISSANO, AND R. LEVI-M• NTALCINI. 1982. Effect of oral administration of nerve growth factor and its antiserum on sympathetic ganglia of neonatal mice. Dev. Brain Res. 4: 31-34.

3. BARBIN, G., I. SELAK, M. MANTHORPE, AND S. VARON. 1984. Use of central neuronal cultures for the detection of neuronotrophic agents. Neuroscience, in press.

4. BJERRE, B., A. BJORKLUND, AND D. CAIRD-EDWARDS. 1976. Axonal regeneration of pe- ripheral adrenergic neurons: effects of antiserum to nerve growth factor in mouse. Ceil Tissue Res. 148: 441-476.

5. BJORKLUND, A., AND U. STENEVI. 1972. Nerve growth factox stimulation of regenerative growth of central noradrenergic neurons. Science 175: 1251-1253.

6. B~TTENSTEIN, J., S. SKAPER, S. VARON, AND G. SATO. 1980. Selective survival of neurons from chick embryo sensory ganglionic dissociates using defined, serum-free supplemented medium. Exp. Cell Res. 125: 183-190.

218 LONG0 ET AL.

7. LINDSAY, R. 1979. Adult rat brain astrocytes support survival of both NGFdependent and NGF-insensitive neurones. Nature (London) 282: 80-82.

8. LINDSAY, R., P. BARBER, M. SHERWOOD, J. ZIMMER, AND G. RAISMAN. 1982. Astrocyte cultures from adult rat brain. Derivation, characterization and neurotrophic properties of pure astroglial cells from corpus callosum. Brain Res. 243: 329-343.

9. LINGO, F., S. SKAPER, M. MANTHORPE, L. WILLIAMS, G. LUNDBORG, AND S. VARON. 1983. Temporal changes of neuronotrophic activities accumulating in vivo within nerve regeneration chambers. Exp. Neural. 81: 756-769.

10. MANTHORPE, M., M. NIET@~AMPEDRO, S. SKAPER, E. LEWIS, G. BARBIN, F. LONGO, C. COTMAN, AND S. VARON. 1983. Neuronotrophic activity in brain wounds in the developing rat. Correlation with implant survival in the wound cavity. Brain Res. 267: 47-56.

11. NIETO-SAMPEDRO, M., E. LEWIS, C. COTMAN, M. MANTHORPE, S. SKAPER, G. BARBIN, F. LQNGO, AND S. VARON. 1982. Brain injury causes time-dependent increase in neu- ronotrophic activity at the lesion site. Science 217: 860-86 1.

12. NIETO-SAMPEDRO, M., M. MANTHORPE, G. BARBIN, L. WILLIAMS, S. VARON, AND C. COTMAN. 1983. Injury-induced neuronotrophic activity in adult brain. Correlation with survival of delayed implants in the wound cavity. J. Neurosci. 3: 2219-2229.

13. ROTHMAN-SCHONFELD, A., L. THAL, S. HOROWITZ, AND R. KATZMAN. 1981. Heart con- ditioned medium promotes central cholinergic regeneration in vivo. Brain Rex 229: 541- 546.

14. STENEVI, U., B. BJERRE, A. BJORKLUND, AND W. MOBLEY. 1974. Effects of localized in- tracerebral injections of nerve growth factor on the regenerative growth of lesioned central noradrenergic neurones. Brain Rex 69: 2 17-234.

15. TEASDALE, G., AND B. JENNETT. 1974. Assessment of impaired consciousness and coma: a practical scale. Lancet ii: 81-84.

16. VARON, S., AND R. ADLER. 198 1. Trophic and specifying factors directed to neuronal cells. Adv. Cell Neurobiol. 2: 115-163.

17. VARON, S., S. SKAPER, G. BARBIN, I. SELAK, AND M. MANTHORPE. 1984. Low molecular weight agents for the survival of cultured CNS neurons. J. Neurosci., in press.

18. WILLIAMS, L., S. SKAPER, I. SELAK, M. MANTHORPE, AND S. VARON. 1984. Tropic factors for CNS neurons in circulating red blood cells. Submitted for publication.

19. YIP, H., AND B. GRAFSTEIN. 1982. Effect of nerve growth factor on regeneration of goldfish optic axons. Brain Res. 238: 329-339.