Pathogenesis of hematogenous bacterial meningitis in rabbits

Transcription

1 Pathogenesis of hematogenous bacterial meningitis in rabbits F. KARL GREGORIUS, M.D., B. LAMAR JOHNSON, JR., M.D., W. EUGENe: STERN, M.D., AND W. JANN BROWN, M.D. Department of Surgery, Division of Neurosurgery, Department of Medicine, Division of Infectious Diseases, and Department of Pathology, Division of Neuropathology, UCLA School of Medicine, Los Angeles, California ~" The authors present data obtained from a series of 27 rabbits studied following intracarotid injection of saline, brain-heart infusion broth, aerobic, or anaerobic bacteria. These data support the hypothesis that injection of cultures of both aerobic and anaerobic organisms by way of the carotid artery disrupts the blood-brain barrier over the cerebral convexities within 15 minutes; however, the inflammatory response and bacterial proliferation occur much more rapidly in the ventricular system. Within 3 hours, the barrier over the convexities is intact, but leukocytes and organisms may be seen in the long cilia of the aqueductal region. A relative failure of leucotaxis over the convexities of the brain is the most likely explanation t~f these preliminary findings in this experimental model. KEY WORDS 9 experimental meningitis 9 fluorescein angiography 9 scanning electron microscope B ACTERIAL meningitis has been produced successfully in experimental animals only by direct inoculation of organisms into the subarachnoid space or by intravenous inoculation of bacteria after either cerebrospinal fluid (CSF) withdrawal or introduction of foreign material into the subarachnoid space? -s'6 Intracarotid injection of Staphylococcus aureus or Clostridium perfringens reliably produces meningitis in the rabbit and allows study of the proliferation of bacteria and the development of the inflammatory response. Clinical experience suggests that meningitis may follow uncomplicated septicemia, although this has not been verified experimentally in animals to our knowledge. Materials Materials and Methods Twenty-seven New Zealand rabbits were studied. Of 10 control rabbits, five received bilateral intracarotid injections of 1.5 to 2.0 cc of 0.9% saline, four the same amount of brain-heart infusion (BHI) broth, and one received no injection. Fourteen rabbits received S. aureus: 11 had 1.5 to organisms in BHI broth injected into each carotid artery, and three received 1.5 to 2 l0 s organisms. Three additional rabbits received 1.5 to 2 lo s C. perfringens in BHI broth. J. Neurosurg. / Volume 45 / November,

2 F. K. Gregorius, et al. TABLE l Summary of each rabbit studied Material Injected Time from Injection Study Method* Sacrifice & Rabbit No. to Sacrifice M F S T Methodt none 1 X )< perfused saline control 2 15 rain X KC1 3 3 hrs X X X KCI 4 4 hrs X KCI 5 5 hrs X X X perfused 6 6 hrs X X perfused BHI~ broth 7 15 rain X KCI 8 4 hrs X X KCI 9 6 hrs X X X perfused 10 12hrs X X X KC1 S. aureus X 10a/cc in BHI broth min X X KCI rain X X KC rain >( X KCI min X X KCI 15 3 hrs X KCI 16 6 hrs X X X perfused 17 6 hrs X X KCI 18 6 hrs X X perfused 19 12hrs X X X KCI hrs X X X perfused hrs )~ died S. aureus X 10g/cc in BHI broth min X X KCI 23 5 hrs )< KCI 24 6 hrs KCI C. perfringens X 109/ce in BHI broth min X KCI 26 4 hrs )< X X KC hrs X X KCI *M = light microscopy; F = fluorescein angiography; S = scanning electron microscopy; T = transmission electron microscopy. tperfused = aortic perfusion of fixative; KCI = sacrificed with potassium chloride. ~BHI = brain-heart infusion. Between 15 minutes and 17 hours after carotid injection, 26 rabbits were sacrificed by intravenous administration of 25% potassium chloride (six control and 13 infected rabbits) or by intraaortic perfusion of 0.12 M phosphate-buffered 1.2% glutaraldehyde and 1% paraformaldehyde (four control and three infected rabbits). One rabbit infected with S. aureus died spontaneously 18 hours after injection. Technique Cultures of CSF were obtained by incising the skin and neck musculature and swabbing the cisterna magna in eight rabbits, five with S. aureus and three with C. perfringens. Sixteen rabbits were given 1 cc of 10% sodium fluorescein in an ear vein 5 minutes prior to sacrifice with KCI. Their brains were removed and viewed under an ultraviolet light to assess the integrity of the blood-brain barrier over the cerebral convexities. These preparations were photographed. All brains were studied with light microscopy and transmission (TEM) or scanning (SEM) electron microscopy or both (Table 1). The brains of four control and 10 infected rabbits were placed in either 10% for- 562 J. Neurosurg. / Volume 45 / November, 1976

3 Experimental hematogenous bacterial meningitis malin for histological study or phosphate buffered glutaraldehyde for TEM or SEM. The remaining six control and seven infected brains were sectioned sagittally; half of each brain was fixed for routine histological study and the other half prepared for SEM or TEM. Hematoxylin and eosin and Gram- Weigert stains were used for sections to be studied by light microscopy. Results Bacteriological Studies Bacterial studies were performed on five rabbits treated with S. aureus and three with C. perfringens. Cultures of CSF were positive in seven of the eight rabbits studied. Four cultures taken 15 minutes after intracarotid injection and three cultures taken 8, 10, and 18 hours after injection were positive. One culture taken 4 hours after the injection of C. perfringens was negative. Fluorescein Studies Control Animals. Two saline and two broth-injected control specimens showed brilliant staining of the blood vessels over both hemispheres. Single rabbits sacrificed 15 minutes after broth injection and 12 hours after saline injection showed a single minute leak of fluorescein from a blood vessel into the subarachnoid space. Infected Animals. Six rabbits were sacrificed 15 minutes after injection (five with S. aureus and one with C. perfringens). One of these animals receiving 108 S. aureus and one receiving 10 ~ S. aureus showed marked bilateral extravasation of fluorescein into the subarachnoid space (Fig. 1). Two animals receiving 108 S. aureus showed only mild extravasation of the fluorochrome. Six other rabbits were sacrificed 3 to 12 hours after injection (four with S. aureus and two with C. perfringens), and in these fluorescein was confined to blood vessels on the surface of the brain. Light Microscopic Study Control Animals. Ten control rabbits were studied by light microscopy (five treated with saline, four with broth, one without injection). With the exception of red blood cells at ependymal surfaces, other pathological changes were not seen in the meningeal vessels, FIG. 1. Ultraviolet photograph of fluorescein distribution in the cerebral hemispheres and dorsal cerebellum of a rabbit killed 15 minutes after intracarotid injection of S. aureus. Note marked extravasation of fluorescein. meninges, or brain parenchyma in nine of these rabbits. The tenth rabbit, sacrificed 4 hours after saline injection, exhibited numerous cortical and subcortical focal collections of lymphocytes and monocytes. Infected Animals. Meningeal vascular alterations consisting of capillary engorgement with neutrophils and a mild exudate of neutrophils and lymphocytes were seen in all 12 animals infected with S. aureus 15 minutes following injection. Gram-positive cocci were seen in eight of nine infected animals from 6 to 18 hours following carotid injection (Fig. 2). Adherence of leukocytes and Gram-posi- FIG. 2. View of lateral ventricular wall in animal given S. aureus and killed 4 hours later. Note red cells and Gram-positive coccus within the ventricular cavity. Gram Weigert, 475. J. Neurosurg. / Volume 45 / November,

4 F. K. Gregorius, et al. tive cocci to ependymal surfaces in the lateral and fourth ventricles and aqueduct and focal loss of cilia were consistently seen as early as 6 hours after injection. Subcortical petechial hemorrhages were seen once, but no microemboli were seen in any infected animals. All three animals infected with C. perfringens had changes similar to those in animals infected with S. aureus, namely capillary engorgement and a ventricular exudate composed of leukocytes and red cells. Organisms could be positively identified in only one animal sacrificed 8 hours after injection. Flo. 3. An SEM view of cilia of the fourth ventricle of broth-injected animal. Note red cell (arrow). X Scanning Electron Microscopy Of the seven animals in the control group studied by SEM, three were treated with saline, three with broth, and one without injection. As described above, all cellular elements were absent from lumina of vessels and the ventricles of animals sacrificed by perfusion. On SEM could be seen fine wavy patterns of micro- and macrocilia in the lateral ventricle, densely matted cilia in the aqueduct, and rows of ciliary tufts in the fourth ventricle (Fig. 3). In two animals sacrificed with KC1, scattered red blood cells FIG. 4. Leukocyte (L) associated with cocci (arrows) in lateral ventricle of a rabbit killed 6 hours after S. aureus injection. Note distorted cilia. X ,4. J. Neurosurg. / Volume 45 / November, 1976

5 Experimental hematogenous bacterial meningitis were seen in the aqueduct and fourth ventricle. Leukocytes were rarely observed. Infected Animals. The six rabbits infected with S. aureus exhibited numerous leukocytes in the lateral ventricle, aqueduct, and fourth ventricle 3 to 17 hours following injection. There were spheres of 1.5 to 3.0 u in diameter identical to S. aureus seen on SEM by others studying the interaction of leukocytes and microorganisms (Fig. 4).' Leukocytes with organisms were to be seen densely clustered in the long, hair-like cilia of the aqueduct in three different animals (Fig. 5). There was focal loss of the normal fine ciliary pattern. The cellular response was much denser in the aqueduct than in the lateral and fourth ventricles. In one specimen, organisms could be seen in the subarachnoid space through a rent in the arachnoid membrane on the cortex of the frontal lobe (Fig. 6). No morphological differences were noted on SEM study between animals sacrificed at 3 and 17 hours following injection. The three animals infected with C. perfringens exhibited similar intraventricular changes. Leukocytes were clustered densely in the long aqueductal cilia and scattered through the lateral and fourth ventricles. A few 0.5 X 4 # rod-like structures associated with leukocytes were seen on the ventricular floor and on subarachnoid surfaces in a single animal sacrificed 8 hours after injection (Fig. 7). These rod-like structures were not seen in control animals nor in rabbits infected with S. aureus. They resemble no neural microstructural elements we are aware of, or circulating blood elements seen in the rabbit brain. Transmission Electron Microscopy Control Animals. Four rabbits in the control group were studied with TEM; of these two were treated with saline, one with broth, and one was without injection. The ependymal surfaces of the lateral, third, and fourth ventricles in one perfused rabbit showed normal ultrastructural patterns at a magnification of 10,000, while the choroid plexus capillaries in three other perfused rabbits showed normal tight junctions and fenestrations at a magnification of 13,500. Infected Animals. Three rabbits infected with S. aureus were studied with TEM. In one rabbit perfused with fixative 6 hours after injection, ultrastructural alterations were seen in the ependyma. The changes included diffuse disorganization of the extracellular space and loss of micro- and macrocilia at a FIG. 5. Field of variably sized probable ieukocytes and organisms adherent to cilia of third ventricle in a rabbit given S. aureus 3 hours before sacrifice. X J. Neurosurg. / Volume 45 / November,

6 F. K. Gregorius, et al. FIG. 6. Organisms (arrow) breaking through subarachnoid membrane to cortical surface adjacent to small vessel (arrow). Same rabbit as Fig. 4. X FIG. 7. Single rod-shaped organism (arrow) adherent to cilia of third ventricle in rabbit killed 8 hours after injection of C. perfringens. X J. Neurosurg. / Volume 45 / November, 1976

7 Experimental hematogenous bacterial meningitis magnification of 10,000. The other two animals exhibited no ependymal changes in the relatively small number of areas we were able to examine by this method. Choroid plexus capillaries in two rabbits sacrificed 6 and 12 hours after injection showed no changes in either tight junctions or fenestrations. In addition, no unusual pinocytotic vesicles or inclusions were noted in the choroid plexus at a magnification of 13,500. Discussion Except in one saline-injected animal, morphological study of controls by all methods showed only minor changes. The chronic meningoencephalitis exhibited in this single rabbit was not seen in any other control or infected animal and was almost certainly present prior to injection. Observations in the animals injected with fluorescein demonstrated that 15 minutes after intracarotid injection of either test organism, temporary disruption of the bloodbrain barrier occurred over the cerebral hemispheres; however, at 3 to 12 hours, the barrier was again intact. Whether this temporary change was due to bacterial toxins or the hypertonicity of the injected cultures or both is unknown? Since both test organisms studied were cultured from CSF within 15 minutes of intracarotid introduction, it is likely that organisms, like the fluorescein, entered the subarachnoid space promptly after injection. Both light microscopic and scanning electron microscopic observations support the proposition that organisms gain access to the ventricular system and multiply in sufficient numbers to be found in the lateral, third, and fourth ventricle within 6 hours of injection. Organisms were associated with a widespread and extensive intraventricular leukocytic response but were only rarely seen in the subarachnoid space after injection. The destruction of cilia and loss of ciliary pattern in the lateral and fourth ventricle were consistently seen on SEM, histological sections, and one TEM study. With SEM long and thickly matted cilia in the aqueduct were shown enmeshing both organisms and leukocytes. Although bacterial proliferation presumably occurred in animals sacrificed after 6 hours, the intraventricular inflammatory response to a few bacteriological forms was intense in all sections studied. This reaction suggests that hematogenous bacterial meningitis in this experimental preparation progresses most rapidly in the aqueduct and ventricular system and develops at a slower rate in the subarachnoid space. Since bacteria are present in the cisterna magna within 15 minutes after their injection into the internal carotid artery at a time when subarachnoid vessels are permeable to fluorescein, the more rapid development of an intense inflammatory response and bacteria in the ventricular system seems paradoxical. This paradox suggests the possibility that leukocytes may fail to migrate through reconstituted capillaries with tight junctions on the cerebral convexity but still pass through fenestrated vessels in the choroid plexus of the ventricular system. Local proliferation of bacteria and migration of leukocytes into the space surrounding the capillaries with early damage to surrounding ependyma is a likely explanation of the observed changes. Increased perfusion of the choroid plexus as a primary event is less likely. References 1. Austrian CR: Experimental meningococcus meningitis. Bull Johns Hopkins Hosp 29: , Belsey MA: Cerebrospinal fluid-glutamic oxalacetic transaminase activity in experimental meningitis, in Hobby GL (ed): Antimicrobial Agents and Chemotherapy Ann Arbor, Michigan: American Society for Microbiology, 1967, pp Harter DH, Petersdorf RG: A consideration of the pathogenesis of bacterial meningitis: review of experimental and clinical studies. Yale J Biol Med 32: , Klainer AS, Betsch C J: Scanning electron microscopy of the attachment of human polymorphonuclear leukocytes to Staphylococcus aureus. J Infect Dis 127: , Pollay M: Effect of hypertonic solutions on the blood-brain barrier. Neurology 25: , Waggener JD: The pathophysioiogy of bacterial meningitis and cerebral abscesses: an anatomical interpretation, in Thompson RA, Green JR (eds): Advances in Neurology, Volume 6: Infectious Diseases of the Central Nervous System. New York: Raven Press, 1974, pp 1-17 Address reprint requests to: F. Karl Gregorius, M.D., Department of Surgery, Division of Neurosurgery, Wadsworth Veterans Administration Hospital, Los Angeles, California J. Neurosurg. / Volume 45 / November,