Ulllted States Patent [19] [11] Patent Number: 5,053,419 Lipton [45] Date of Patent: Oct. 1, 1991

[54] TREATMENT OF AIDS DEMENTIA, [56] _ References Cited MYELOPATHY AND BLINDNESS PUBLICATIONS [75] Inventor: Stuart A. Lipton, Newton, Mass. Bfennem?n ei 3L, 1988, Nature, 3351639 Price et al., 1988, Science, 239:586. [73] Assignee: The Children’s Medical Center H0 et al., 1989, Annals in Internal Medicine, 111:400. Corporation, Boston, Mass. Pomerantz et al., 1987, New Eng. J. Med., 317:1643. [21] AppL No‘: 502 296 Primary Examiner-Stanley J. Friedman 22 F] d M 30 1990 [57] ABSTRACT [ 1 1c ' an ’ A method of reducing death of CNS neurons in a human patient infected with a human immunode? Related US. Application Data ciency virus, involving administering to the patient a [63] Continuation-impart of Ser. No. 331,872, Mar. 31, compound Capable of rcducing the gplzo'responsive ‘1939, abandoned‘ rise in free Ca+ + ion concentration in the CNS neurons of the patient, in a concentration effective to cause such [51] Int. Cl.5 ...... A61K 31/44 reduction. [52] US. Cl...... , ...... 514/356 [58] Field of Search ...... 514/356 10 Claims, 6 Drawing Sheets US. Patent Oct. 1, 1991 Sheet 1 of 6 5,053,419

Surviving retinal ganglion cells

-e rgp120-3B *- gp120-RF2

-~q.4a-.

-1 3 ~12 40 log concentration (M)

Surviving retinal ganglion cells E] solitary Q clustered

- F|G.2 US. Patent 0¢t.1,1991' - sheet 2 of 6 5,053,419

[JA\\\\\\\\1 ? 3-01 ° 200 pM gp120 I ' 20 pM gp120 25; = 2pM gp120 1 J 2.0 l [Ca 2"1, (pM) 1.5% 1.0% 0.5

[Ca2+], (PM)

0 - 260 400 660 Concentration (pM) FIG.3B US. Patent 0a. 1,1991 Sheet 3 of 6 5,053,419

1.5- _"F_

1.0

y [ca2*], (PM) -1\‘ 05- \ Wm§ m Control gp120‘ gp120+ gp120+ pre post immune immune

FIG.4

3.0 - ° 200 pM gp120 I I 200 pM gp120 + 100 nM 2:5 -

2.0 :

1.0:

0.5 " US. Patent 0a. 1, 1991 Sheet 4 of 6 5,053,419

[1 solitary @ clustered

oO00O -quanna4l1¢4 O826.4.

1 T Cont-fol.///////////////////////////// gp120-3B nlfedipineT%/////////////////¢ gp120+ (1 0 FM) nlfedipineTW//////////////////% w///////////////// control 'gp120-RFr2' nifedlplne ‘ 0+ (10 PM) nl e ipine

FIG.5B US. Patent 0a. 1, 1991 Sheet 5 of 6 5,053,419 11111 8

r .r t l’ gp120-3Br%////////% nif edipine 9 20+ (10PM) nie i ine FIG.6A

ed

2D% 4

. Fm.. AM n 8 g + (10 PM) nife i ine U.S. Patent Oct. 1, 1991 Sheet 6 of 6 5,053,419

200pA VH=-40mV \ 50ms VC=-1OmV cell #481 #481

“PM -40 -30 -20 -10 10 20 30 —¢ 1 : 1 i 1 Wm")

plus 400

-200 q FIG.8B gp120-RF2 I -3oo

-5OO - 5,053,419 1 2

TREATMENT OF AIDS DEMENTIA, Drawings MYELOPATHY AND BLINDNESS FIG. 1 is a dose response curve showing retinal gan glion cell death at different gpl20 concentrations; BACKGROUND OF THE INVENTION 5 FIG. 2 is a graph of retinal cell survival in the pres This application is a continuation in part of U.S.S.N. ence of recombinant gpl20 and antiserum; FIG. 3(a) is a graph of kinetics of intracellular free 331,872, ?led Mar. 31, 1989, now abandoned. This invention relates to the treatment of central Ca2+concentration ([Ca2+]i) in a retinal ganglion cell in response to various doses of gpl20, and (b) is a steady nervous system disorders caused by infection with state dose response graph of gpl20 concentration versus human immunode?ciency virus type I (HIV-I). [Ca2+]; HIV-I infection in humans causes general immuno FIG. 4 is a bar graph of gpl20 promoted rise in suppression and involves other disorders, such as blind [Ca2+]in the presence of gpl2O alone or gpl20 that has ness, myelopathy, or a dementing neurological disorder, been immunoprecipitated with gplZO antiserum; i.e., the AIDS dementia complex, the latter of which is 15 FIG. 5 is a graph of retinal cell survival in the pres a common and important cause of morbidity in patients ence of gpl20 and/or lOuM nifedipine; in advanced stages of infection. HIV-1 infection has FIG. 6 is a graph of retinal cell survival in the pres been documented in various areas of the CNS, including ence of gpl20 and/or lpMM nifedipine or nimodipine; the cerebral crotex, spinal cord, and retina. Price et al. FIG. 7 is a graph of intracellular [Ca2+]in the pres (1988, Science 239:586) and Ho et a1. (1989, Annals in ence of gpl20 and the Ca2+channel antagonist nimodi Internal Medicine 111:400) review the clinical, epidemi pine; ological, and pathological aspects of the AIDS demen FIG. 8(a) shows current flowing through calcium tia complex, and suggest that the mechanism underlying channels in the presence of nifedipine and gpl20, and the neurological dysfunction may be indirect tissue (b) is a graph of the current-voltage relationship. damage by either viral- or cellular-derived toxic sub 25 stances released by infected cells. SELECTION OF CA+ +CHANNEL Pomerantz et al. (1987, New Eng. J. Med. 271:1643) ANTAGONIST document the presence of HIV type I infection of the Any suitable antagonist, generally, of neuronal volt retina in two patients with AIDS. Brenneman et a1. age~dependent Ca+ +channels may be used to reduce or (1988, Nature 3352639) found gpl20, the coat‘protein of prevent AIDS related vision loss, myelopathy, or de HIV, killed hippocampal neurons. mentia. Preferred calcium channel antagonists include, SUMMARY OF THE INVENTION but are not limited to, the following drugs, of which the most preferred are those that are capable of crossing the The invention features a method of reducing death of blood brain barrier, for example, nimodipine (Miles CNS neurons in a human patient infected with a human 35 Pharmaceuticals, West Haven, CT) Smith Kline drug immunode?ciency virus, by administering to the patient no. 9512 (Smith Kline, French Beecham, Philadelphia, a compound capable of reducing the gpl20 responsive PA), and diproteverine (Smith, Kline, French-Bee rise in intracellular free Ca+ +ion concentration in CNS cham) . Less preferred antagonists are those that are less neurons of the patient, in a concentration effective to CNS permeable, for example, (Calan, G.D. cause such reduction. Searle & Co., Chicago, Ill.; Isoptin, Knoll, Whippany, Preferably, the blood of the infected patient contains NJ), , (Cardizem, Marion, Kansas antibodies to HIV 1; most preferably, the patient mani City, MO), and nifedipine, U.S. Pat. No. 3,485,847, fests symptoms of AIDS related complex or of acquired hereby incorporated by reference (Procardia, L P?zer, immune de?ciency syndrome. The method may also NY, NY; Adalat, Miles). Other Ca2+channe1 antago include administration of a second compound that is 45 nists which may be useful are mioflazine, , capable of reducing the gpl20 responsive rise in intra , , CERM-l96, R 58735, R-56865, cellular free Ca++ion concentration. The compounds , , , PNZOO-llO, most preferred are calcium channel antagonists capable , , R-(—)-202-79l, and R-(+) of crossing the blood brain barrier; these may be admin Bay K-8644 (Miles, Bayer), whose chemical formulae istered orally or intravenously; of these compounds, are described in Boddeke et al., Trends in Pharmaco nimodipine is currently most preferred. Calcium chan logic Sciences (1989) 10:397 and Triggle et al., Trends nel antagonists that do not freely cross the blood brain in pharmacologic Sciences (1989) 10:370. barrier are less preferred; these may be administered For any given calcium channel antagonist, effective intrathecally to the brain and/ or spinal cord, or intravit ness in preventing neurological disorders associated really to the retina; most preferred among these antago 55 with HIV-1 (or other HIV) infection is determined by nists are nifedipine, verapamil, nitrendipine, diltiazem, screening the drug using one or more of the following nicardipine, flunarizine, and diproteverine; oral or in assays of neuronal cell function; i.e., neuronal cell traveous routes of administration may be effective. death, detection of intracellular free Ca2+ion concen The invention can reduce dementia, myelopathy, or tration in neurons, and detection of current flow vision loss associated with infection by a human im through Ca2+channels An effective antagonist will munode?ciency virus. cause a decrease in HIV-l-associated neuronal cell Other features and advantages of the invention will death, and will prevent the rise in intracellular Ca2+ion be apparent from the following description of the pre concentration that occurs in the presence of gpl20. In ferred embodiments thereof, and from the claims. addition, an effective antagonist will decrease Ca+ rion 65 in?ux through neuronal calcium channels to a degree DESCRIPTION OF THE PREFERRED suf?cient to reduce neuronal cell death, while not com EMBODIMENTS pletely blocking Ca++ion in?ux, an event which itself The drawings will ?rst briefly be described. might kill neuronal cells. The antagonist may be com 5,053,419 3 4 pounded into a pharmaceutical preparation, using phar signal sequence (Berman et al., 1985, Science 227:1490) maceutical compounds well-known in the art; the exact to allow the envelope protein to be constituitively se formulation of the anagonist compound depends upon creted by the CH0 cell line. Production in a mamma the route of administration. lian cell ensured that the envelope protein was glycosy lated. This envelope glycoprotein, rgpl20-3B, was puri ASSAYS FOR NEURONAL CELL FUNCTION ?ed by immunoaf?nity chromatography to 5 parts in a AND DEATH million (99.995%) pure based on estimates from poly An antagonist may be tested for utility in the method acrylamide gel electrophoresis and Western blotting. of the invention using any type of neuronal cell from the The preparations of gpl20 (at low concentrations) were central nervous system, using the following assays, as highly labile in that they had to be freshly thawed (with long as the cell can be isolated intact using conventional refreezing avoided) in order to display activity. techniques. Retinal cultures were used in the following Cell survival was assayed after one day in culture. assays (but hippocampal cortex neurons have also been Incubations lasted 20-24 h at 37° C. in an atmosphere of used, e.g., in assays of neuronal death and intracellular 5% CO2/95% air. Ganglion cells could be unequivo calcium), because they can be produced from postnatal cally identified by the continued presence of the ?uores mammals, are well-characterized, and contain a central cent blue dye. The ability of retinal ganglion cells to neuron, the retinal ganglion cell, that can be unequivo take up and cleave ?uorescein diacetate to ?uorescein cally identi?ed with ?uorescent labels. A substantial was used as an index of their viability as described in portion of retinal ganglion cells in culture display both detail in Hahn et al., 1988, supra. Dye uptake and cleav functional synaptic activity and bear many, if not all, of age correlates well with normal electrophysiological the neurotransmitter receptors found in the intact re properties assayed with patch electrodes. tina. To perform the viability test, the cell-culture medium was exchanged for physiological saline containing GP120 INCREASES NEURONAL CELL DEATH 0.0005% ?uorescein diacetate for 15-45 5, and then cells IN VITRO 25 were rinsed in saline. Retinal ganglion cells that did not Neuronal cell death was assayed by incubating retinal contain the ?uorescein dye (and thus were not living) ganglion cells in vitro with puri?ed native or recombi often remained visible under both phase contrast and nant gplZO and scoring live cells. The ability of the UV ?uorescence optics, the latter because of the contin Ca2+channel antagonist to reduce neuronal cell death ued presence of the marker dye granular blue; other was determined by scoring live cells which had been dead retinal ganglion cells had disintegrated and only incubated with both gpl20 and the antagonist. debris remained. In contrast, the viable retinal ganglion Retinal ganglion cells from postnatal rats were identi cells displayed not only a blue color in the UV light but ?ed and their viability ascertained as follows. Under also a yellow-green fluorescence with ?lters appropri general anesthesia, the fluorescent dye granular blue ate for ?uorescein. Thus, the use of two exchangeable (Mackrornolekulare Chemic, Umstadt, FRG) was in fluorescence ?lter sets permitted the rapid determina jected as approximately a 2% (w/v) suspension in saline tion of viable ganglion cells in the cultures, which were into the superior colliculus of 4- to 7-day-o1d Long found as solitary neurons or lying among other cells in Evans rats (Charles River Laboratory, Wilmington, small clusters (usually in the ratio of approximately 1:10 MA). Two to 7 days later, the animals were killed by solitary to clustered). decapitation and enucleated, and the retinas quickly 40 FIG. 1 is a dose response curve for concentrations of removed. The retinas were dissociated and cultured in gpl20 ranging from 10*9 M to less than 10-13 M, and Eaqle‘s minimum essential medium (MEM, catalog shows that incubation of native puri?ed gp120 (RFZ) or #1090, Gibco Grand Island, NY), supplemented with recombinant gpl20 preparations with cultured retinal 0.7% (w/v) methylcellulose, 2 mM glutamine, 1 ug/ml cells resulted in the death of a signi?cant number of gentamicin, l6mM dextrose, and 5%(v/v) rat serum, as ganglion cells within 24 h. A signi?cant increase described in Lipton et al., 1987, J. Physiol. 385:361. The (P<0.0l) in cell death was observed at gpl20 concen cells were then plated onto 75 mm2 glass coverslips trations above 2>2 uM) were obtained with concentra triplicate or quadruplicate samples), 10 uM nifedipine tions of gpl20 at or above 200 pM. induced signi?cant killing on its own compared to con The increases in [Ca2+] observed with gpl20 could trols; in fact, in all six experiments, gp120 + nifedipine still have been caused by a contaminant in the puri?ed treated culture dishes had somewhat higher levels of preparation of the viral envelope protein, although this 20 survival than that observed with nifedipine alone seems unlikely with the highly puri?ed recombinant (P<0.05, Sign test). One possible explanation for these gpl20. As additional confirmation, we performed im phenomena invokes the hypothesis that there is an opti munoprecipitation experiments with goat antibody to mal level of intracellular Ca2+necessary for neuronal gpl20 (anti gplZO) coupled to protein A coated Se health and welfare. Too little Caz'l'may inhibit survival pharose beads. 25 (as seen here in the presence of nifedipine) while too Immunoprecipitation of gpl20 was performed as much Ca2+may also lead to cell death (as evidenced by described above with some modi?cations. A 1:100 0r the effect of gplZO on Caz'tcurrent) In between these 1:500 dilution of anti gpl20 or preimmune serum from two extremes (the “control” value of Ca2+, or the the same goat, was bound to protein A coated Se Ca2+concentration in the presence of gplZO +nifedi pharose beads, washed, and incubated with a solution pine), survival may be enhanced. Alternatively, nifedi containing 7 nM gplZO for 18 hours at 4° C.; this was pine alone could be toxic to these neurons for some followed by centrifugation. The supernatant of the ma other, unrelated reason, although in that case it would terial treated with preimmune serum had gpl20 activity be dif?cult to explain the ?nding that survival was as evidence by immunoblotting and by producing an slightly better with the combination of gplZO plus 10 increase in [Ca2+] and cell death (after a dilution of 35 uM nifedipine compared to 10 uM nifedipine alone. lz35O to ~20 pM); the material exposed to anti gpl20 The effect of using lower doses of calcium channel had little or no activity. antagonists on neuronal survival following gplZO treat One of three such experiments is shown in FIG. 4. ment is shown in FIG. 6. In FIG. 6a, a lower dosage Treatment with preimmune serum did not signi?cantly (luM) of nifedipine antagonizes the lethal effect of alter the ability of gpl20 to increase [Ca2+]i (compare 40 gpl20 (20 pM) without substantial depression of retinal striped and grey bars; although the mean [Ca2+]i was ganglion cell survival by the calcium channel antagonist greater after application of gpl20 treated with preim itself. Treated cultures received gpl20-3B (20 pM) and mune serum than after gpl20 alone, this difference did /or nifedipine (1 uM) at the time of plating, and retinal ' not reach statistical signi?cance. Both gpl20 and preim ganglion cell survival was assayed one day later. Incu mune serum-treated gplZO produced a signi?cant in 45 bation with gplZO resulted in neuronal cell killing com crease in [Ca2+compared to the control (P<0.0l, anal pared to control sibling cultures (P <0.01). The pres ysis of variance (ANOVA) followed by Scheffe multi ence of 1 uM nifedipine alone had substantially less ple comparison of means; signi?cance indicated by an detrimental influence on retinal ganglion cell survival asterisk). In contrast, immunoprecipitation with postim than 10 uM (as observed in FIG. 5). Therefore, by ?ne mune serum containing anti-gpl20 completely abro tuning the dose response curve, it is possible to ?nd an gated the gplZO effect (last column on right of FIG.). optimal level of calcium channel antagonist that pro ' duces minimal death on its own and yet substantially GPl20-PROMOTED NEURONAL CELL DEATH blocks the toxicity produced by gpl20. IS PREVENTED BY CALCIUM CHANNEL Therapeutic options for decreasing neuronal cell ANTAGONISTS 55 death due to HIV-I infection should include the use of Nifedipine (SOOnM - luM) antagonizes the lethal other speci?c calcium channel antagonists, e. g., nimodi effect of native gpl20 isolates 38 and RF2 on retinal pine, which has better CNS permeability than nifedi~ ganglion cells. In FIG. 5, cultures received gpl20 pine (Allen et al., 1983, New Eng. J. Med. 308:619). (ZOpM) and/or nifedipine (10 pM) at the time of plating, FIG. 6b shows that treatment with 1 uM nimodipine and neuronal cell survival was assayed one day later. signi?cantly saved clustered retinal ganglion cells from Each value is the mean obtained in 4 replicate tissue gpl20-induced death (P<0.0l); however, 1 uM nimodi culture dishes. Survival of clustered retinal ganglion pine produced killing by itself similar to that observed cells in the control group was signi?cantly greater than with 10 uM nifedipine (FIG. 5). FIG. 6a shows that those treated with gpl20 (P<0.0l). Compared to gpl20 lowering the concentration of nifedipine to 1 uM also treated cultures, the addition of nifedipine signi?cantly 65 resulted in a decrease in gplZO-induced cell death; and, increased retinal ganglion cell survival (P<0.01). In at this concentration, treatment with nifedipine alone comparison to the controls, exposure to 10 uM nifedi did not result in neuronal cell killing. Taken together, pine alone resulted in cell death (P<0.05), although it the results with nifedipine and nimodipine demonstrate 5,053,419 9 10 that the dosage of calcium channel blockers is obviously prove effective in different areas of the brain. For exam an important parameter; too high a concentration ap ple, the Ca2+current in hippocampal neurons has been pears to harm normal neurons in the absence of gpl20. shown to be partially suppressed by verapamil (100 pM) FIG. 7 shows the effect of lOO nM nimodipine on the (Yaari et al., 1987, Science 235:680); in addition, novel intracellular concentration of free Ca2+ions in retinal calcium channel blockers or G proteins and intracellu ganglion cells incubated with gpl20. The cells were lar messengers that affect their efficacy, may prove loaded with fura 2, as described above. The viral enve useful -in this regard (Olivera et al., 1985, Science lope protein gpl20 (200 pM) was applied by puffer 2302338, Dolphin et al., 1987 J. Physiol. 386zl; and Yaari pipette to neurons previously bathed in normal medium et al., 1987, Science 238:1288). Miller (suora) showed or in medium containing 100 nM to 1 uM nimodipine that hippocampal neurons were more sensitive to cal for several min. In FIG. 7, for neurons bathed in normal cium channel antagonists than were striatal neurons. calcium medium (0; n=l0), gpl20 produced an in This variability can probably be attributed to the fact crease in [Ca2+] as expected. In contrast, after treat that only a prolonged component of Ca2+current (simi ment with nimodipine (10; n: l l), gpl20 did‘not elicit a lar to L-type current) is sensitive to dihydropyridines. signi?cant increase in [Ca2+]. 15 USE CURRENT-FLOW IN THE PRESENCE OF GPlZO IS INHIBITED BY CALCIUM CHANNEL Antagonists of the invention may be administered by ANTAGONISTS any one of a number of routes; for example, for a CNS permeable antagonist such as nimodipine or Smith The following assay of neuronal cell function tests Kline drug no. 9512, nicardipine, flunarizine, or di the effect of calcium channel antagonists on Ca2+ion proteverine, administration may be oral or intravenous, flow through Caz‘l'channels. Without being bound to at a divided dosage of 60 to 1200 mg/day. For less CNS any theory as to the mechanism whereby gpl20 in permeable antagonists, administration may be intrathe creases cell death, it is possible that gpl20 increases cally or intravitreally, but oral or intravenous adminis current flow across Ca1+channels As a precautionary 25 tration may be effective especially in the presence of measure in screening for a compound capable of reduc ing the gpl20 associated rise in intracellular Ca2+con break-down of the blood brain barrier in sick patients; centration, the following assay of Ca2+current should nifedipine is adminstered at a dosage of 20 to 800 be performed in the presence of gpl20. mg/day; verapamil is administered at a dosage of 80 to FIG. 8 shows that nifedipine suppresses Ca2+cur 720 mg/day; and diltiazem is administered at a dosage rents in the presence of gpl20 in retinal ganglion cells. of 60 to 960 mg/day. In FIG. 8a, in another neuron, nifedipine (10 pM) par I claim: tially antagonized the current carried by Ba2+through 1. A method of reducing death of CNS neurons in a calcium channels during the application of 20 pM human patient infected with a human immunodefi gpl20~RF2. All three traces were obtained in the pres ciency virus, comprising administering to said patient a ence of gpl20. Inhibition by nifedipine was especially compound capable of reducing the gpl20 responsive apparent (53% decrease) when the voltage step (in this rise in free Ca+ +ion concentration in said CNS neurons case, VC=—lO mV) was initiated from a depolarized of said patient, in a concentration effective to cause such holding potential (V 5:40 mV); this observation is reduction. consistent with the action of calcium channel blockers 2. The method of claim 1 wherein the blood of said on Ca2+current in other tissues. FIG. 8b shows the I-V human patient contains antibodies to HIV 1. relationship for the same cell as in FIG. 8a, and illus 3. The method of claim 1 wherein said human patient trates the effect of nifedipine at various test potentials in manifests symptoms of the AIDS related complex or the continued presence of 20 pM gplZO RF2. Recording acquired immune deficiency syndrome. conditions use barium as the charge carrying species. A 45 4. The method of claim 1, said compound being capa similar concentration range for antagonism of Ca2+cur ble of crossing the blood brain barrier. rent with nifedipine (100 nM-lOuM) was obtained with 5. The method of claim 1, said compound being nimodipine (lOOnM-lOuM). nimodipine or Smith Kline drug no. 9512. 6. The method of claim 1, said nimodipine being ad OTHER NEURONS ARE SENSITIVE TO GP12O ministered to said patient orally or intravenously. Another important consideration is whether or not 7. The method of claim 1, said compound being ad Ca2+1evels in mammalian central neurons other than ministered to said patient intrathecally or intravitreally. retinal ganglion cells will be sensitive to gpl20 and, 8. The method of claim 1, said compound being one therefore, amenable to treatment with calcium channel or more of nifedipine, verapamil, nitrendipine, or dil blockers. Miller (1987, Science 235:46) suggests that 55 tiazem. dihydropyridines affect Ca2+influx, at least to some 9. The method of claim 5, in which said compound is degree, in over 90% of neurons from various brain nimodipine. areas; there were, however, regional differences in the 10. The method of claim 1, said compound being effectiveness of these drugs. Other classes of calcium nicardipine, flunarizine, or diproteverine. channel antagonists, such as phenylalkylamines, may # =8 ! i i

65