fournal of Vimlogical ~ethads, 33 ( 199 1) K-100 0 1991 Elsevier Science Publishers B.V. / 0168-8510/91/$03.50
87
A~~~~*~OI6885109I~l89J VIRMET 01173
A semiautomated multiparameter approach for anti-HIV drug screening Robert J. Gulakowski’, James B. McMahon~, Patricia G. Staley2, Robert A. Moran’ and Michael R. Boyd’ ‘Laboratory of Drug Discovery Research and Development, NCI-FCRDC, Frederick Cancer Research and Development Center. Frederick, Maryland, U.S.A. and 2Program Resources, Inc., Frederick, Maryland, U.S.A.
(Accepted 14February 1991)
Summary We are implementing a series of complementary assays for initial follow-up confirmation and prioritization of new active anti-HIV compounds identified by the U.S. National Cancer Institute’s large-scale in vitro primary anti-HIV screen. Two different kinds of cellular viability assays, in addition to specific assays for total cellular DNA content, supematant reverse transcriptase activity, p24 core antigen production and the synthesis of infectious HIV virions are all performed from a single well of a 96well microtiter plate containing human host cells infected with HIV. Antiviral activities of several known prototype HIV inhibitors including 3’-azido,3’-deoxythymidine, 2’,3’-dideoxycytidine, dextran sulfate and phorbol myristate acetate were compared in these multiparameter assays as a means of validation. Procedures to automate the method optimally, as well as to maximize the safety of the technicians working with HIV and HIV-infected cells have been emphasized. The resulting semiautomated, highly reproducible battery of assays yields a maximum amount of antiviral and cytotoxicity information from a minimum amount of sample. This is especially crucial when analyzing new synthetic compounds and natural product extracts orfractions where the available amounts of sample may be very limited. Anti-HIV assay; Fluorescence;
Correspondence
Drug screening
to: Dr. James B. McMahon.
21702%1201,U.S.A.
NCI-FCRDC, Bldg. 1052, Room 121, Frederick, MD.
88
The present study was undertaken to develop safe and automatable ‘Stage II’ methods to complement the large-scale, primary (‘Stage I’) anti-HIV screen operated by the U.S. National Cancer Institute (NCI) (Boyd, 1988). Stage II screening assays as described here are intended to facilitate the initial follow-up confirmation andprioritization of new active leads identified by the primary screen. The NCI’s Stage I antiHIV screening assay is based upon the cellular reduction of the tetrazolium salt, XTT, to a soluble, colored formazan which is quantitated by spectrophotometry (W&low et al., 1989). Human lymphoblastoid target cells infected and killed by HIV reduce little or no XTT; prevention of HIV’s cytopathic effects by an active anti-HIV substance is detected by the mainten~ce or enhancement of XTT reduction. In the approach now described for secondary evaluation of newly-identified active leads, a battery of interrelated assays on individual wells from 96-well microtiter plates are carried out concurrently. Cellular viability is estimated by both an adaptation of the XTT method above, and a method using the fluorescent probe 2’-7’biscarboxyethyl-5(6)-carboxyfluorescein acetoxymethyl ester (BCECF) (Rink et al., 1982). BCECF is a nonfluorescenL molecule which readily enters viable cells where it is hydrolyzed by cellular esterases to a fluorescent molecule. Total cellular DNA content is measured with the dye, 2-diamidino-phenylindole (DAPI), which fluoresces when intercalated at A-T-specific sites in chromatin (McCaffrey et al., 1988). These dyes are used in combination with the Particle concentration fluorescent immunoassay (PCFIA), specifically the Screen MachineTM developed by Baxter Healthcare Coloration (Mundelein, IL). The Screen Machine is a semiautomated fluorescentplate reader capable of adding reagents and/or wash buffers to filter-bottomed, 96well plates with the subsequent evacuation of fluid and concentration of fluorescently stained cells on the cellulose acetate filter. Fluorescence is then detected via epifluorescence. Concurrent with the above, confirmatory assays of p24 antigen production, reverse transcriptase activity and synthesis of infectious virions are also performed. Assay modifications have been made to increase automation and to address critical safety considerations for working with HIV and HIV-infected cells. Studies of the effects of selected prototype antiviral agents on HIV-infected T-lymphoblastoid cells in vitro support the view that these multiparameter assays can facilitate the initial confirmation, characterization and comparative evaluations of potential new anti-HIV agents and provide the basis for their prioritization for more detailed investigation and/or drug development.
Materials and Methods Cells and vkus
The lymphocytic cell line CEM-SS was obtained from Dr. Peter Nara of the National Cancer Institute-Frederick Cancer Research and Development Center and maintained in RPM1 1640 medium (Gibco, Grand Island, NY) without phenol red and
89
supplemented with 5% fetal bovine serum (FBS) (Gibco), 2 mM t-glutamine and 50 &g/ml gentamicin (Gibco) (complete medium). exponentially growing CEM-SS cells were pelleted and resuspended at a concentration of 2.0 x lo5 cells/ml in complete medium. The Haitian variant of HIV, HTLV-IIIRF [3.54 x IO6 syncytium forming units (SFU)/ml] which is being used for large-scale screening by the NCI, was used throughout the study. Frozen virus stock solutions were thawed immediately before use and resuspended in complete medium to yield 1.2 x 10” SFU/ml.
All experimental antiviral agents including the tetrazolium reagent, XTT (Weislow et al., 1989), were obtained from the Drug Synthesis and Chemistry Branch, Developmental Therapeutics Program, Division of Cancer Treatment, National Cancer Institute. Crystalline stock materials were stored at -70°C and solubilized in 100% dimethylsulfoxide (DMSO). Drugs were diluted in complete medium with the final concentration of DMSO not exceeding 1% (v/v). Compounds used in these studies included 3’-azido,3’-deoxythymidine (AZT; NSC 602670), 2’,3’- dideoxycytidine (ddC; NSC 606170), dextran sulfate (NSC 620255) and phorbol myristate acetate (PMA; NSC 262244). Biscarboxyethyl-5(6)-carboxyfluorescein acetoxymethyl ester (BCECF) was purchased from Molecular Probes, Inc. (Eugene, OR) and dissolved immediately before use in DMSO (1 mg/ml). A working solution of 2 fig/ml was prepared in Dulbecco’s phosphate-buffered saline (PBS) (Gibco). 4’,6-diamidino-2-phenylindole (DAPI) was purchased from Sigma Chemical Co. (St. Louis, MO). Stock solutions of DAPI were prepared at lOO~g/ml in distilled water by sonication, passed through a0.45m filter and stored at -20°C. Working solutions of DAPI were prepared at 10 pg/ml in PBS containing 0.5% nonidet P-40 (NP-40) (Sigma). XTT was prepared at a concentration of 1 mg/ml in serum-free RPM1 1640. Phenazine methosulfate (PMS) (Sigma) was prepared at 0.153 mg/ml in PBS and stored at -20°C. Immediately before use, XTT was dissolved at 37°C and PMS was added to yield a final concentration of 20 PM* Anti-HIV assays All serial drug dilutions, reagent additions and plate-to-plate transfers were carried out using the automated Biomek 1000 Workstation (Beckman Instruments, Palo Alto, CA). Each experimental antiviral agent was initially diluted in complete medium and added to a single column of a 96-we11 microtiter plate (dilution plate). The Biomek was used to perform eight serial dilutions of each drug and to transfer a X00-,&aliquot of each dilution to the test plate. Uninfected CEM-SS cells were plated at a density of 1 x 104cells in 50~1 of complete medium. Diluted virus was then added to appropriate wells in a volume of 50 pl to yield a multiplicity of infection of 0.6. Appropriate cell, virus and drug controls were used with the final volume in each well being 200,&. Fig. 1 is a schematic of the plate format that has been developed for the assay procedures. Uninfested, untreated cell controls (C) and untreated virus-infected cell controls (V)
90
LOW-
DRUG DlLUrlON
+
H(GH
A
Fig. I. Schematic diagram of 96well plate format. V: virLls-infected. untreated ceils: C: uninfected, untreated cells.
are placed on both sides of the test plate and drug blanks are placed along the top and bottom of the plate. Cells that receive test compounds are included in quadruplicate virus-infected wells and duplicate uninfected wells. This format was developed in order to maximize automated drug and cell transfers and to take into account the higher variability in the virus-infected cell assay endpoints. Plates were then incubated at 37°C in an atmosphere containing 5% CO? for 6 days. A flow diagram of the assay is shown in Fig. 2. After incubation, aliquots of cell-free supernatant were removed from each well using the Biomek, and analyzed for reverse transcriptase activity, ~24 antigen production and synthesis of infectious virions. A 25-~1 sample of 0.002% (w/v) Fluoricon reference particles (590/620 nm) (Baxter Healthcare Corp.) was added to each well of the test plate to be used as an internal standard for fluorescence assays. The Biomek was then used to evenly disperse the contents of each well of the test plate and transfer 50-,LJ aliquots to each of two new microtiter plates. These plates were then used to measure either ceilular viability using BCECF or total DNA content using DAPI. X7T assay Toestimate cellular viability, the metabolic reduction of the tetrazolium salt, XTT, to the soluble, colored formazan was carried out by adding .50,~1of the XTT/‘PMS solution to each well of the original test plate (Fig. 2) and incubating for4 h at 37°C. After incubation the plates were covered with adhesive plate sealers (Dynatech, Alexandria, VA), shaken and their optical densities determined using a V-max photometer (Molecular Devices, Inc., Menlo Park, CA) at a test wavelength of 450 nm.
91
1. Remove 50 ~1 supernatant a. b.
2. Reference particles
RT activity p24antigen
production
c. infectious virions
Mlcrotlter well 200 /II
1 0 W 3. Resuspend cells
5. xl-r WI
I
4. Transfer 50 ~1 allquots 5. BCECF
5. DAPI
6. Transfer to Pandex plate
6. Transfer to Pandex plate
485/535nm
450 nm
4W450 nm
7. Read Fig. 2. Schematic diagram of the
BCECF
assay procedures.
assay
Cellular viability was also measured using BCECF. Freshly prepared BCECF solution (25 ~1) was added to each well of the microtiter plate (Fig. 1). These plates were incubated at 37°C for 30 min. After incubation, 25 ,~l of a 2% solution of parafoirnaldehyde (final concentration = 0.05%) was added to each well in order to inactivate the virus. Incubation with the paraformaldehyde continued for 30 min. The contents of each well were mixed and a 75~1 aliquot was then transferred to a filter-bottomed 96well plate (Baxter Healthcare Corp.). The BCECF plate was then placed in the Screen
92
Machine programmed to execute the following protocol: add 20 ~1 of a 0.25% w/v suspension of 3.2-pm polystyrene beads (Baxter Healthcare Corp.) in PBS as a filtration support matrix; filter away the liquid phase using a vacuum pressure of 15 mm Hg for 1‘/2 min; wash the cell-bead cake in each well with PBS using a vacuum pressure of 20 mm Hg for 1 min; read fluorescence of each well (signal channel = excitation at 485 nm, emission at 535 nm and reference channel = excitation at 590 nm, emission at 620 nm). DAFT assay Total DNA content of each well was determined by the following modifications to the method described by McCaffrey et al. (1988). The contents of each well were fixed by adding 25 ~1 of a 2% paraformaldehyde solution and incubating the plate at 37°C for 30 min. 25 ,ul of the DAPI/NP-40 solution (Fig. 1) was then added to each well and incubated for 2 h. The contents of each well were mixed and a 75~1 aliquot was transferred to a filter-bottomed 96well plate (Baxter Healthcare Corp.). The DAPI plate was then placed in the Screen Machine and processed by the same protocol as the BCECF plate above with the signal channel set at an excitation of 400 nm and an emission of 450 nm. ~24 assay The production of the HIV- 1 internal core p24 antigen was measured using a p24 antigen-capture assay (Coulter Immunology, Hialeah, FL). Supernatants from test plates were diluted 1: 100 in 10% Triton X-100 and stored frozen at -20°C until needed. Two-hundred-microliter aliquots of Triton X-treated samples were added to microtiter wells previously coated with a murine monoclonal anti-HIV p24 antigen. The plate was sealed and incubated at 37°C for 1 h. Plate washings were carried out using an automated Denley Wellwash 4 (Coulter Immunology) plate washer. After washing and blotting dry the plate, 200 ~1 of a biotinylated human monoclonal antiHIV- 1 p24 was added to appropriate wells, and the plates were reincubated for 1 h at 37°C. After additiona washing, 200 ~1 of a streptavidin-horseradish peroxidase solution was added and the plate was incubated for 30 min at 37°C. A tetramethyIbenzene solution was added to each well and incubated at room temperature for 30 min. Following incubation, an acidic stopping reagent was added to each well and the absorbance read at 450 nm within 30 min using a Vmax photometer (Molecular Devices). The concentration of p24 was determined by comparison with a standard curve of known p24 concentrations. Syncytium assay The syncytium assay described by Nara et al. (1987) was used for quantitation of infectious virus. Supe~atants from test plates were examined in CEM-SS cell monolayers at multiple dilutions to obtain countable levels of SFU (5~200 SFU/well) in Z4 days.
93
Reverse t~ansc~~pt~seassay A 304 aliquot of supernatant was added to 30 ,~l of a virus disruption buffer containing 50 mM Tris, pH 7.8,0.15 mg/ml dithiothreitol (DTT) andO.l% Triton X-100. A lO-~1 sample of lysed virus was added to 30 ,~l of a cocktail containing 2 ,~l of 1 M Tris, pH 7.8, 1 ,~l of 3 M KCl, 5 ,ul of 3 mg/ml DTT, 5 ~1 of 0.1 M magnesium acetate, 10~1 of Poly(rA).p(dT),o (2 units/ml) (Pharmacia, Piscataway, NJ), 6.5 ,ul of distilled HZO, 0.5 ,~l of 10% TritonX-lOOand 10~1 of [“HJdTTP (16.56Ci/mmol) (Amersham Corp., Arlington Heights, IL). Samples were incubated for 30 min at 37”C, harvested on to DE8 1 ion-exchange paper and allowed to adsorb for 15 min. Sample pads were rinsed six times with 5% Na2HP04 and then twice with distilled H20. Pads were then dried and counted in a liquid scintillation counter. Samples were counted in triplicate.
Results Linearity of assay endpoint to cell number Exponentially growing CEM-SS cells were harvested, washed and plated in 96well microtiter wells at varying cell concentrations. Following their inoculation, the cells were treated with either XTT, BCECF or DAPI according to the protocols outlined in Materials and Methods. The linearity of response of each of the three endpoints to cell input is shown in Fig. 3. The fluorescence assays, using BCECF and DAPI, (Fig. 3A) showed excellent linearity over a wide range of cell concentrations. Reproducible results could be obtained from both assays with cell numbers below 1000 cells
lo2
+
DAPI
m
BCECF
(400/450
nm)
(485/535
+Il?lqlq lo2
nm)
0.0 -1 lo3
10' Cell Number
lo5
10"
0
25
50
Cell Number
75
100
(x 1000)
Fig. 3. Linearity of cellular assays. (A) fluorescence of CEM-SS cells (RFU f SEM) following staining with DAPI (+) or BCECF (m) and (B) XTT-formazan production [O.D. i SEM) by CEM-SS cells (A).
125
94
per assay. The calorimetric XTT assay (Fig. 3B) also showed good linearity but with a higher detection limit of 5000-10 000 cells per assay. Analysis of the esfects of anti-/W
compounds on acute HIV infections
The effects of several known anti-HIV compounds were dete~ined on acutely infected cells using the multiparameter assay described. The results using the clini-
A)
BCECF 120
8)
(Viability)
DAPI
a
P “; 2 .E
60 40
-
: Lt
120
-
100
-
0 >g
80-
P z L .c
60
-
40
-
20
-
(DNA
Content)
@=&&e-o
‘&
: 20
-
Cl
bp
XTT (Metabolic
‘J)
Activity)
0 3 c 2 TI u ii z! +
: 20
-
BO60-
+ A
RT activity p24 production n infectious virus
40-
w
u
0.001
0.01
0.1
0.001
10
1
0.01
0.1
1
10
[A=] (PM) 0
Uninfected
@
Infected
CEM-SS CEM-SS
Cells Cells
Fig. 4. Effect of AZT on uninfected (0) and HIV-infected (0) CEM-SS cells assessed after6 days in culture using (A) cellular viability, (B) total DNA content, (C) metabolic activity, (D) supematant reverse transcriptase activity (+), p24 antigen production (A) and synthesis of infectious virions (W). Values for infected cultures represent the mean ofquadruplicate samples (SEM
XTT (Metabolic 160
5
Activity)
D>
-
140-
z L r ::
125 100
+ RT activity A p24 production n infectious virus
.-
as
25 0
0 l
Uninfected GEM-SS Ceils Infected CEM-SS Cells
Fig. 5. Effect of ddC on uninfected (0) and HIV-infected(e) CEM-SS cells assessed after 6 days in culture using (A) cellular viability, (R) total DNA content, (C) metabolic activity, (D) supematant reverse transcriptase activity (a), p24 antigen production (A) and synthesis of infectious virions (a). Values for infected cultures represent the mean of quadruplicate samples (SEM 5 8%) while uninfected values represent the mean of duplicate samples. All points are graphically represented as percentages of the uninfected, non-drug treated control values.
96
A)
BCECF
6)
(Viability)
OAPI (DNA c0nt.w) 120
z : z 1 s z ._ : bp
cl
XTT (Metabolic 120
2 ‘; d P 2
looBO60 4020 -
D>
Activity)
7
175 150
looBO-
.’ -
.‘;
80 40-
: w
20 -
o0.001
.’ .’ _.
0.01
0.1
-
[Dextran 0 0
125
” z I
100
E 0-
50
iR
25
A-A
Sulfate]
+ RT activity A p24 production n infectious virus
\\ r
75 i
0 1 0.001
10
1
k.:
1
0.01
0.1
1
10
(kg/ml)
Uninfected CEM-SS Cells Infected CEM-SS Cells
Fig. 6. Effect of dextran sulfate on uninfected (0) and HIV-infected (0) CEM-SS cells assessed after 6 days in culture using (A) cellular viability,(B) total DNA content, (C)metabolic activity,(D) supematant reverse transcriptase activity (+), p24 antigen production (A) and synthesis of infectious virions (W). Values for infected cultures represent the mean of quadruplicate samples (SEM I So/,), while uninfected values represent the mean of duplicate samples. All points are graphically represented as percentages of the uninfected, non-drug-treated control values.
acute HIV infection. However, it was also evident that ddC was less potent than AZT in this assay. In addition, the results indicated that under the in vitro conditions employed, ddC had more adverse cellular effects than AZT. For example, at 10 PM ddC, there was an inhibitory effect on viability (Fig. 5A) and DNA content (Fig. 5B). Interestingly, this was not shown by an inhibition of the metabolic capacity measured by XTT-formazan production (Fig. 5C). Dextran sulfate, a known inhibitor of HIV infections in vitro, was evaluated under conditions identical to those of the nucleoside analogs. The results are shown in Fig. 6. Dextran sulfate protected CEM-SS cells against HIV-induced toxicity and infectious virus release as demonstrated by all parameters assayed (Fig. 6A-D).
97
A)
BCECF
DAPI
(Viability)
(DNA
Content)
140
z z x zD z .: x 2
100 60 60 40 7.0 120 I
cl
XTT 160 o : z n u 0 % z : .bp
(Metabolic
Activity) 400
-
140120-
/\
loo60 60 40 20
-
L&
/
-
o10-6
RT activity p24 production n infectious virus
$
300
”z
250
$ g
2oo 150
5 r
100 50
z 10-S
+ A
350
/
10-4
10 -3
0 i
1o-2
0
Uninfected
.
Infected
CEM-SS CEM-SS
Cells Cells
Fig. 7. Effect of PMA on uninfected (0) and HIV-infected (0) CEM-SS cells assessed after 6 days in culture using (A) cellular viability, (B) total DNA content, (C) metabolic activity (D) supematant reverse transcriptase activity (+), p24 antigen production (A) and synthesis of infectious virions (m). Values for infected cultures represent the mean of quadruplicate samples (SEM $8%), while uninfected values represent the mean of duplicate samples. All points are graphically represented as percentages of the uninfected, non-drug-treated control values.
The potential utility of multiparameter evaluations of anti-HIV compounds was most strikingly exemplified by results obtained with PMA (Fig. 7). Other studies have shown that PMA inhibits HIV-induced syncytium formation (Chowdhury et al., 1990) while enhancing virus production. Identical results were obtained in this study. While protecting CEM-SS cells from the cytopathic effects of HIV, PMA caused a marked decrease in DNA content (Fig. 7B) and cellular metabolic activity (XTT) of both uninfected and infected cultures (Fig. 7C). The BCECF viability assay revealed that this inhibitory effect was due primarily to a decreased rate of proliferation (cytostasis) rather than to overt cytotoxicity (Fig. 7A). Unlike the other four com-
98
pounds examined in this study, PMA appeared to have a biphasic response on virus production in the infected cultures. At lower concentrations, there was a marked increase in the production of infectious virus particles and viral antigens (Fig. 7D). At higher concentrations, however, virus production by infected cells was markedly inhibited.
Discussion The quest for new compounds, synthetic or naturally derived, with anti-HIV activity has led numerous investigators to search for rapid, reproducible, quantifiable and safe screening techniques. The ideal anti-HIV screening system would provide info~ation on multiple effects of test drugs on both the virus and host cell. Monitoring the direct effects of drugs on host test cells is important in that it allows for the normalization of detected antiviral effects with the number and viability of test cells. A variety of anti-HIV screening assays have been described which involve a single endpoint or a combination of endpoints. The metabolic reduction of the tetrazolium salt, MTT, has been used singly to evaluate anti-HIV compounds (Pauwels et al., 1988) or in combination with a supematant reverse transcriptase assay (Schwartz et al., 1988). Fluorescein diacetate (FDA) has also been used to evaluate cellular viability for both HIV-infected MT-4 cells (Schols et al., 1988) as well as human tumor cells (Larsson et al., 1989). As described in the introduction, the XTT-tetrazolium assay (Weislow et al., 1989) is currently used by the NC1 for the initial large-scale Stage I screening of synthetic chemicals and natural products for anti-HIV activity. The screen has been used to evaluate tens of thousands of compounds for anti-HIV activity and has effectively identified several new classes of antiviral compounds. The use of the XTT method offers advantages over the MTT procedure for large-scale anti-HIV screening from a biosafety standpoint as it requires fewer pipetting, aspiration or centrifugation steps. Past screening experience has shown that the XTT method, as with any colorimetric procedure, has a disadvantage when testing colored compounds such as the suramin analog dyes and many natural product extracts. Therefore, in our multiparameter battery of assays we employ a second, complementary assay of cellular viability to detect drug-treated HIV-infected cells using the carboxyfluorescein derivative BCECF. BCECF has recently been used to effectively determine the viability of drug-treated peripheral blood lymphocytes (Leeder et al., 1989) and was found to be more sensitive than the MTT assay. BCECF was chosen for use in our system for its linearity of fluorescence detection with very low numbers of cells (Fig. 3a) and also because its fluorescence was not quenched during fixation of HIVinfected cells with paraformaldehyde. Total cellular DNA content is measured using the A-T intercalating dye, DAPI. DAPI has previously been established as an accurate means to measure cell density of bovine and rat smooth muscle cells in vitro (McCaffrey et al., 1988). Averett ( 1989) developed an assay system, using the Screen Machine, to evaluate anti-HIV compounds using the nucleic acid dye propidium iodide (PI) to measure cytoprotec6on in combination with an antigen capture assay to measure HIV core p24 protein. DAPI
99
was chosen to measure total DNA content in this study because of its increased signal detection with small numbers of cells (Fig. 3) as compared with PI (Averett, 1989). In addition, cells stained with DAPI retain their fluorescence for long periods of time and can be stored for further analyses such as cell size comparisons. Certain classes of compounds being screened for anti-HIV activity may interfere directly with the endpoint analysis of any of the calorimetric, immunological, fluorescent and/orradioactive assays and therefore may give either a false positive or negative result. With this in mind it is clear that the best option is to obtain the most cellular and viral info~ation as possible to be able to compare and provide the broadest activity profile for candidate drugs. In our strategy for Stage II screening of new compounds we determine the relative effeets of known and potential antiviral compounds by the simultaneous assessment of cellular viability, metabolic activity and DNA content of virally infected, drug-treated lymphoblastoid cells, as well as measuring drug effects on infectious virus release, p24 antigen production and reverse transcriptase activity. Results obtained in the present study, using several known prototype inhibitors of HIV-induced cytopathology, establish the validity and suggest the general usefulness of these assays for initial follow-up confirmation and comparison of new compounds. AZT exerted a broad effect against acute HIV infections as indicated by the return to uninfe~tedcontrol cell values for viability, metabolic and DNAcontent assays (Fig. 4A-C) as well as the dramatic decrease of viral activity (Fig. 4D). The antiviral activity of ddC (Fig. 5) was also well documented in the multiparameter assay. However, at aconcentration of IO,UM, ddC caused adecrease in cellular viability (Fig. 5A) and DNA content (Fig. 5B) in both uninfected and infected cells. Similar results were seen with ddC-treated H9 cells chronically infected with HTLV-IIIB (data not shown). The ability of dextran sulfate to protect CEM-SS cells from the cytopathic effects of HIV in an in-well infection was demonstrated by all assay parameters (Fig. 6). Similar results were obtained when dextran sulfate was added to previously (bulk) infected CEM-SS cells (data not shown). These results are in agreement with a recent report which demonstrated that dextran sulfate can block both HIV adsorption and syncytium formation (Montefiore et al., 1990). The dete~ination of the direct effects of an antiviral agent on the target cells is important in the selection and overall evaluation of candidate compounds for drug development. For example, notwithstanding its tumor promoting properties, phorbol myristic acetate (PMA) has other properties which, if observed in isolation, might encourage its consideration as a potential anti-HIV agent. While PMA clearly protected GEM-SS cells from the cytopathic effects of HIV, there was also a dramatic direct effect of PMA on the cells themselves (Fig. 7B). As clearly apparent in the multiparameter analysis, the lower concentrations of PMA gave signs of protection as evident by an increase in viability (Fig. 7A) and metabolic activity (Fig. 7C). However, PMA appeared to stimulate the production and release of infectious virus and related viral proteins (Fig. 7D). In addition, PMA at moderate concentrations exerted strong growth inhibitor effects on both infected and uninfected cells. The assays described here are performed in conjunction with a Biomek 1000T~~ Workstation (Beckman), a semiautomated system for drug dilution, reagent addition
100
and plate-to-plate transfers and can be used for all steps including the addition of cells and virus solutions to test plates. In addition, the fhtorescence assays are performed using the Screen ~a~hineTM (Baxter Healthcare Corp.), a semiautomated fluorescent plate reader which has been described elsewhere (Averett, 1989; Leederet al., 1989). These semiautomated systems are ideal for work with HIV-infected cells to minimize direct technician involvement, to decrease the possibility of human error and more importantly to decrease possible exposure of staff to the lethal pathogen. The assays described provide a safe, semiautomated~ and reproducible muItip~ameter approach to screening compounds for antiviral activity. These assays provide a maximum amount of information with a minimum of sample, which is a critical issue when screening naturally-derived products or fractions thereof, where the amount of sample is often iimited. In addition, this multiparameter assay may aid in the prioritization of potential anti-HIV compounds for drug development. References Averett, D.R. (1989) Anti-HIV compound assessment by two novel high capacity assays. J. Virol. Methods 23,263-216. Boyd, M.R. f 1988) Strategies for the identification ofnew agents for the treatment of AIDS: a national program to facilitate the discovery and preclinical development of new drug candidates Forclinical evaluation. In: V.T. DeVita, S. Hellman and S.A. Rosenberg (Edsj. AIDS Etiology, Diagnosis, Treatment and Prevention, Lippincott, Philadelphia, pp. 305-3 17. Chowdhury, I.H., Koyanagi. Y., Kobayashi, S., Hamamoto, Y.. Yoshiyama, H, Yoshida, T. and Yamamoto, N. (1990) The phorbol ester TPA strongly inhibits HIV-l-induced syncytia formation but enhances virus production: possible involvement of protein kinase C pathway. Virology 176,126-l 32. Larsson, R. and Nygren, P. (1989) A rapid fluorometric method for semjautomated dete~ination ofcytotoxicity and cellular proliferation of human tumor cell lines in microculture. Anticancer Res. 9, 11I I-
i 120. Leeder, J.S., Dosch, H.M., Harper, P.A., Lam, P and Spielberg, S.P, (1989) Fluorescence-based viability assay for studies of reactive drug intermediates. Anal. B&hem. 177,3&l-372. McCaffrey, T.A., Agarwal, L.A. and Weksler, B.B. (1988) A rapid fluorometric DNA Assay for the measurement of cell density and proliferation in vitro. In Vitro Cellular Dev. Biol. 24,247-252. Montefiori, DC., Robinson, W.E., Modliszewski, A., Rowland, J.M., Schuffman, S.S. and Mitchell, W.M. (1990) Differential inhibition of HIV- I cell binding and HIV- I -induced syncytium fomlation by low molecular weight sulphated polysaccharides. J. Antimicrob. Chemother. 25.313-3 IX. Nara, P.L., Hatch, W.C., Dunlop, N.M., Robey, W.G., Arthur, L.O., Gonda, M.A. and Fischinger, P.J. (1987) Simple, rapid,quantitative, syncytium forming microassay forthe detection ofhuman immunodeficiency virus neutralizing antibody. AIDS Res. Hum. Retroviruses 3.2833302. Pauwels, R., Balzarini, J., Baba, M., Snoeck, R., Schols, D., Herdewijn, P., Desmyter, J. and De Clercq, E. (1988) Rapid and automated tetrazofium-based calorimetric assay for the detection of anti-HIV compounds. J. Virol. Methods 20.309-321. Rink, T.J., Tsien. R.Y. and Pozzan, T. (1982) Cytoplasmic pH and free Mg” in lymphocytes. J. Cell. Biol. 95.189-196. Schols, D., Pauwels, R., Vanla~~endonck, F., Balzarini, J. and De Clercy, E. ( 198X)A highly reliable, sensitive, flow cytometric/~uorometric assay for the evaluation of the anti-HIV activity of antiviral compounds in MT-4 cells. J. Immunol. Methods I 14,27-32. Schwartz, O., Henin, Y., Mare&al, V. and Montagnier, I,. (1988) A rapid and simple calorimetric test for the study of anti-HIV agents. AIDS Res. Hum. Retroviruses 14,441-44X. Weislow, OS., Riser, R., Fine, D.L., Bader, J.. Shoemaker, R.H. and Boyd, M.R. (19891 New soluhle-formazan assay for HIV-l cytopathic effects: application to high-flux screening of synthetic and natural products for AIDS-antiviral activity. J. Natl. Cancer Inst. 8 I, 577-586.
87
A~~~~*~OI6885109I~l89J VIRMET 01173
A semiautomated multiparameter approach for anti-HIV drug screening Robert J. Gulakowski’, James B. McMahon~, Patricia G. Staley2, Robert A. Moran’ and Michael R. Boyd’ ‘Laboratory of Drug Discovery Research and Development, NCI-FCRDC, Frederick Cancer Research and Development Center. Frederick, Maryland, U.S.A. and 2Program Resources, Inc., Frederick, Maryland, U.S.A.
(Accepted 14February 1991)
Summary We are implementing a series of complementary assays for initial follow-up confirmation and prioritization of new active anti-HIV compounds identified by the U.S. National Cancer Institute’s large-scale in vitro primary anti-HIV screen. Two different kinds of cellular viability assays, in addition to specific assays for total cellular DNA content, supematant reverse transcriptase activity, p24 core antigen production and the synthesis of infectious HIV virions are all performed from a single well of a 96well microtiter plate containing human host cells infected with HIV. Antiviral activities of several known prototype HIV inhibitors including 3’-azido,3’-deoxythymidine, 2’,3’-dideoxycytidine, dextran sulfate and phorbol myristate acetate were compared in these multiparameter assays as a means of validation. Procedures to automate the method optimally, as well as to maximize the safety of the technicians working with HIV and HIV-infected cells have been emphasized. The resulting semiautomated, highly reproducible battery of assays yields a maximum amount of antiviral and cytotoxicity information from a minimum amount of sample. This is especially crucial when analyzing new synthetic compounds and natural product extracts orfractions where the available amounts of sample may be very limited. Anti-HIV assay; Fluorescence;
Correspondence
Drug screening
to: Dr. James B. McMahon.
21702%1201,U.S.A.
NCI-FCRDC, Bldg. 1052, Room 121, Frederick, MD.
88
The present study was undertaken to develop safe and automatable ‘Stage II’ methods to complement the large-scale, primary (‘Stage I’) anti-HIV screen operated by the U.S. National Cancer Institute (NCI) (Boyd, 1988). Stage II screening assays as described here are intended to facilitate the initial follow-up confirmation andprioritization of new active leads identified by the primary screen. The NCI’s Stage I antiHIV screening assay is based upon the cellular reduction of the tetrazolium salt, XTT, to a soluble, colored formazan which is quantitated by spectrophotometry (W&low et al., 1989). Human lymphoblastoid target cells infected and killed by HIV reduce little or no XTT; prevention of HIV’s cytopathic effects by an active anti-HIV substance is detected by the mainten~ce or enhancement of XTT reduction. In the approach now described for secondary evaluation of newly-identified active leads, a battery of interrelated assays on individual wells from 96-well microtiter plates are carried out concurrently. Cellular viability is estimated by both an adaptation of the XTT method above, and a method using the fluorescent probe 2’-7’biscarboxyethyl-5(6)-carboxyfluorescein acetoxymethyl ester (BCECF) (Rink et al., 1982). BCECF is a nonfluorescenL molecule which readily enters viable cells where it is hydrolyzed by cellular esterases to a fluorescent molecule. Total cellular DNA content is measured with the dye, 2-diamidino-phenylindole (DAPI), which fluoresces when intercalated at A-T-specific sites in chromatin (McCaffrey et al., 1988). These dyes are used in combination with the Particle concentration fluorescent immunoassay (PCFIA), specifically the Screen MachineTM developed by Baxter Healthcare Coloration (Mundelein, IL). The Screen Machine is a semiautomated fluorescentplate reader capable of adding reagents and/or wash buffers to filter-bottomed, 96well plates with the subsequent evacuation of fluid and concentration of fluorescently stained cells on the cellulose acetate filter. Fluorescence is then detected via epifluorescence. Concurrent with the above, confirmatory assays of p24 antigen production, reverse transcriptase activity and synthesis of infectious virions are also performed. Assay modifications have been made to increase automation and to address critical safety considerations for working with HIV and HIV-infected cells. Studies of the effects of selected prototype antiviral agents on HIV-infected T-lymphoblastoid cells in vitro support the view that these multiparameter assays can facilitate the initial confirmation, characterization and comparative evaluations of potential new anti-HIV agents and provide the basis for their prioritization for more detailed investigation and/or drug development.
Materials and Methods Cells and vkus
The lymphocytic cell line CEM-SS was obtained from Dr. Peter Nara of the National Cancer Institute-Frederick Cancer Research and Development Center and maintained in RPM1 1640 medium (Gibco, Grand Island, NY) without phenol red and
89
supplemented with 5% fetal bovine serum (FBS) (Gibco), 2 mM t-glutamine and 50 &g/ml gentamicin (Gibco) (complete medium). exponentially growing CEM-SS cells were pelleted and resuspended at a concentration of 2.0 x lo5 cells/ml in complete medium. The Haitian variant of HIV, HTLV-IIIRF [3.54 x IO6 syncytium forming units (SFU)/ml] which is being used for large-scale screening by the NCI, was used throughout the study. Frozen virus stock solutions were thawed immediately before use and resuspended in complete medium to yield 1.2 x 10” SFU/ml.
All experimental antiviral agents including the tetrazolium reagent, XTT (Weislow et al., 1989), were obtained from the Drug Synthesis and Chemistry Branch, Developmental Therapeutics Program, Division of Cancer Treatment, National Cancer Institute. Crystalline stock materials were stored at -70°C and solubilized in 100% dimethylsulfoxide (DMSO). Drugs were diluted in complete medium with the final concentration of DMSO not exceeding 1% (v/v). Compounds used in these studies included 3’-azido,3’-deoxythymidine (AZT; NSC 602670), 2’,3’- dideoxycytidine (ddC; NSC 606170), dextran sulfate (NSC 620255) and phorbol myristate acetate (PMA; NSC 262244). Biscarboxyethyl-5(6)-carboxyfluorescein acetoxymethyl ester (BCECF) was purchased from Molecular Probes, Inc. (Eugene, OR) and dissolved immediately before use in DMSO (1 mg/ml). A working solution of 2 fig/ml was prepared in Dulbecco’s phosphate-buffered saline (PBS) (Gibco). 4’,6-diamidino-2-phenylindole (DAPI) was purchased from Sigma Chemical Co. (St. Louis, MO). Stock solutions of DAPI were prepared at lOO~g/ml in distilled water by sonication, passed through a0.45m filter and stored at -20°C. Working solutions of DAPI were prepared at 10 pg/ml in PBS containing 0.5% nonidet P-40 (NP-40) (Sigma). XTT was prepared at a concentration of 1 mg/ml in serum-free RPM1 1640. Phenazine methosulfate (PMS) (Sigma) was prepared at 0.153 mg/ml in PBS and stored at -20°C. Immediately before use, XTT was dissolved at 37°C and PMS was added to yield a final concentration of 20 PM* Anti-HIV assays All serial drug dilutions, reagent additions and plate-to-plate transfers were carried out using the automated Biomek 1000 Workstation (Beckman Instruments, Palo Alto, CA). Each experimental antiviral agent was initially diluted in complete medium and added to a single column of a 96-we11 microtiter plate (dilution plate). The Biomek was used to perform eight serial dilutions of each drug and to transfer a X00-,&aliquot of each dilution to the test plate. Uninfected CEM-SS cells were plated at a density of 1 x 104cells in 50~1 of complete medium. Diluted virus was then added to appropriate wells in a volume of 50 pl to yield a multiplicity of infection of 0.6. Appropriate cell, virus and drug controls were used with the final volume in each well being 200,&. Fig. 1 is a schematic of the plate format that has been developed for the assay procedures. Uninfested, untreated cell controls (C) and untreated virus-infected cell controls (V)
90
LOW-
DRUG DlLUrlON
+
H(GH
A
Fig. I. Schematic diagram of 96well plate format. V: virLls-infected. untreated ceils: C: uninfected, untreated cells.
are placed on both sides of the test plate and drug blanks are placed along the top and bottom of the plate. Cells that receive test compounds are included in quadruplicate virus-infected wells and duplicate uninfected wells. This format was developed in order to maximize automated drug and cell transfers and to take into account the higher variability in the virus-infected cell assay endpoints. Plates were then incubated at 37°C in an atmosphere containing 5% CO? for 6 days. A flow diagram of the assay is shown in Fig. 2. After incubation, aliquots of cell-free supernatant were removed from each well using the Biomek, and analyzed for reverse transcriptase activity, ~24 antigen production and synthesis of infectious virions. A 25-~1 sample of 0.002% (w/v) Fluoricon reference particles (590/620 nm) (Baxter Healthcare Corp.) was added to each well of the test plate to be used as an internal standard for fluorescence assays. The Biomek was then used to evenly disperse the contents of each well of the test plate and transfer 50-,LJ aliquots to each of two new microtiter plates. These plates were then used to measure either ceilular viability using BCECF or total DNA content using DAPI. X7T assay Toestimate cellular viability, the metabolic reduction of the tetrazolium salt, XTT, to the soluble, colored formazan was carried out by adding .50,~1of the XTT/‘PMS solution to each well of the original test plate (Fig. 2) and incubating for4 h at 37°C. After incubation the plates were covered with adhesive plate sealers (Dynatech, Alexandria, VA), shaken and their optical densities determined using a V-max photometer (Molecular Devices, Inc., Menlo Park, CA) at a test wavelength of 450 nm.
91
1. Remove 50 ~1 supernatant a. b.
2. Reference particles
RT activity p24antigen
production
c. infectious virions
Mlcrotlter well 200 /II
1 0 W 3. Resuspend cells
5. xl-r WI
I
4. Transfer 50 ~1 allquots 5. BCECF
5. DAPI
6. Transfer to Pandex plate
6. Transfer to Pandex plate
485/535nm
450 nm
4W450 nm
7. Read Fig. 2. Schematic diagram of the
BCECF
assay procedures.
assay
Cellular viability was also measured using BCECF. Freshly prepared BCECF solution (25 ~1) was added to each well of the microtiter plate (Fig. 1). These plates were incubated at 37°C for 30 min. After incubation, 25 ,~l of a 2% solution of parafoirnaldehyde (final concentration = 0.05%) was added to each well in order to inactivate the virus. Incubation with the paraformaldehyde continued for 30 min. The contents of each well were mixed and a 75~1 aliquot was then transferred to a filter-bottomed 96well plate (Baxter Healthcare Corp.). The BCECF plate was then placed in the Screen
92
Machine programmed to execute the following protocol: add 20 ~1 of a 0.25% w/v suspension of 3.2-pm polystyrene beads (Baxter Healthcare Corp.) in PBS as a filtration support matrix; filter away the liquid phase using a vacuum pressure of 15 mm Hg for 1‘/2 min; wash the cell-bead cake in each well with PBS using a vacuum pressure of 20 mm Hg for 1 min; read fluorescence of each well (signal channel = excitation at 485 nm, emission at 535 nm and reference channel = excitation at 590 nm, emission at 620 nm). DAFT assay Total DNA content of each well was determined by the following modifications to the method described by McCaffrey et al. (1988). The contents of each well were fixed by adding 25 ~1 of a 2% paraformaldehyde solution and incubating the plate at 37°C for 30 min. 25 ,ul of the DAPI/NP-40 solution (Fig. 1) was then added to each well and incubated for 2 h. The contents of each well were mixed and a 75~1 aliquot was transferred to a filter-bottomed 96well plate (Baxter Healthcare Corp.). The DAPI plate was then placed in the Screen Machine and processed by the same protocol as the BCECF plate above with the signal channel set at an excitation of 400 nm and an emission of 450 nm. ~24 assay The production of the HIV- 1 internal core p24 antigen was measured using a p24 antigen-capture assay (Coulter Immunology, Hialeah, FL). Supernatants from test plates were diluted 1: 100 in 10% Triton X-100 and stored frozen at -20°C until needed. Two-hundred-microliter aliquots of Triton X-treated samples were added to microtiter wells previously coated with a murine monoclonal anti-HIV p24 antigen. The plate was sealed and incubated at 37°C for 1 h. Plate washings were carried out using an automated Denley Wellwash 4 (Coulter Immunology) plate washer. After washing and blotting dry the plate, 200 ~1 of a biotinylated human monoclonal antiHIV- 1 p24 was added to appropriate wells, and the plates were reincubated for 1 h at 37°C. After additiona washing, 200 ~1 of a streptavidin-horseradish peroxidase solution was added and the plate was incubated for 30 min at 37°C. A tetramethyIbenzene solution was added to each well and incubated at room temperature for 30 min. Following incubation, an acidic stopping reagent was added to each well and the absorbance read at 450 nm within 30 min using a Vmax photometer (Molecular Devices). The concentration of p24 was determined by comparison with a standard curve of known p24 concentrations. Syncytium assay The syncytium assay described by Nara et al. (1987) was used for quantitation of infectious virus. Supe~atants from test plates were examined in CEM-SS cell monolayers at multiple dilutions to obtain countable levels of SFU (5~200 SFU/well) in Z4 days.
93
Reverse t~ansc~~pt~seassay A 304 aliquot of supernatant was added to 30 ,~l of a virus disruption buffer containing 50 mM Tris, pH 7.8,0.15 mg/ml dithiothreitol (DTT) andO.l% Triton X-100. A lO-~1 sample of lysed virus was added to 30 ,~l of a cocktail containing 2 ,~l of 1 M Tris, pH 7.8, 1 ,~l of 3 M KCl, 5 ,ul of 3 mg/ml DTT, 5 ~1 of 0.1 M magnesium acetate, 10~1 of Poly(rA).p(dT),o (2 units/ml) (Pharmacia, Piscataway, NJ), 6.5 ,ul of distilled HZO, 0.5 ,~l of 10% TritonX-lOOand 10~1 of [“HJdTTP (16.56Ci/mmol) (Amersham Corp., Arlington Heights, IL). Samples were incubated for 30 min at 37”C, harvested on to DE8 1 ion-exchange paper and allowed to adsorb for 15 min. Sample pads were rinsed six times with 5% Na2HP04 and then twice with distilled H20. Pads were then dried and counted in a liquid scintillation counter. Samples were counted in triplicate.
Results Linearity of assay endpoint to cell number Exponentially growing CEM-SS cells were harvested, washed and plated in 96well microtiter wells at varying cell concentrations. Following their inoculation, the cells were treated with either XTT, BCECF or DAPI according to the protocols outlined in Materials and Methods. The linearity of response of each of the three endpoints to cell input is shown in Fig. 3. The fluorescence assays, using BCECF and DAPI, (Fig. 3A) showed excellent linearity over a wide range of cell concentrations. Reproducible results could be obtained from both assays with cell numbers below 1000 cells
lo2
+
DAPI
m
BCECF
(400/450
nm)
(485/535
+Il?lqlq lo2
nm)
0.0 -1 lo3
10' Cell Number
lo5
10"
0
25
50
Cell Number
75
100
(x 1000)
Fig. 3. Linearity of cellular assays. (A) fluorescence of CEM-SS cells (RFU f SEM) following staining with DAPI (+) or BCECF (m) and (B) XTT-formazan production [O.D. i SEM) by CEM-SS cells (A).
125
94
per assay. The calorimetric XTT assay (Fig. 3B) also showed good linearity but with a higher detection limit of 5000-10 000 cells per assay. Analysis of the esfects of anti-/W
compounds on acute HIV infections
The effects of several known anti-HIV compounds were dete~ined on acutely infected cells using the multiparameter assay described. The results using the clini-
A)
BCECF 120
8)
(Viability)
DAPI
a
P “; 2 .E
60 40
-
: Lt
120
-
100
-
0 >g
80-
P z L .c
60
-
40
-
20
-
(DNA
Content)
@=&&e-o
‘&
: 20
-
Cl
bp
XTT (Metabolic
‘J)
Activity)
0 3 c 2 TI u ii z! +
: 20
-
BO60-
+ A
RT activity p24 production n infectious virus
40-
w
u
0.001
0.01
0.1
0.001
10
1
0.01
0.1
1
10
[A=] (PM) 0
Uninfected
@
Infected
CEM-SS CEM-SS
Cells Cells
Fig. 4. Effect of AZT on uninfected (0) and HIV-infected (0) CEM-SS cells assessed after6 days in culture using (A) cellular viability, (B) total DNA content, (C) metabolic activity, (D) supematant reverse transcriptase activity (+), p24 antigen production (A) and synthesis of infectious virions (W). Values for infected cultures represent the mean ofquadruplicate samples (SEM
XTT (Metabolic 160
5
Activity)
D>
-
140-
z L r ::
125 100
+ RT activity A p24 production n infectious virus
.-
as
25 0
0 l
Uninfected GEM-SS Ceils Infected CEM-SS Cells
Fig. 5. Effect of ddC on uninfected (0) and HIV-infected(e) CEM-SS cells assessed after 6 days in culture using (A) cellular viability, (R) total DNA content, (C) metabolic activity, (D) supematant reverse transcriptase activity (a), p24 antigen production (A) and synthesis of infectious virions (a). Values for infected cultures represent the mean of quadruplicate samples (SEM 5 8%) while uninfected values represent the mean of duplicate samples. All points are graphically represented as percentages of the uninfected, non-drug treated control values.
96
A)
BCECF
6)
(Viability)
OAPI (DNA c0nt.w) 120
z : z 1 s z ._ : bp
cl
XTT (Metabolic 120
2 ‘; d P 2
looBO60 4020 -
D>
Activity)
7
175 150
looBO-
.’ -
.‘;
80 40-
: w
20 -
o0.001
.’ .’ _.
0.01
0.1
-
[Dextran 0 0
125
” z I
100
E 0-
50
iR
25
A-A
Sulfate]
+ RT activity A p24 production n infectious virus
\\ r
75 i
0 1 0.001
10
1
k.:
1
0.01
0.1
1
10
(kg/ml)
Uninfected CEM-SS Cells Infected CEM-SS Cells
Fig. 6. Effect of dextran sulfate on uninfected (0) and HIV-infected (0) CEM-SS cells assessed after 6 days in culture using (A) cellular viability,(B) total DNA content, (C)metabolic activity,(D) supematant reverse transcriptase activity (+), p24 antigen production (A) and synthesis of infectious virions (W). Values for infected cultures represent the mean of quadruplicate samples (SEM I So/,), while uninfected values represent the mean of duplicate samples. All points are graphically represented as percentages of the uninfected, non-drug-treated control values.
acute HIV infection. However, it was also evident that ddC was less potent than AZT in this assay. In addition, the results indicated that under the in vitro conditions employed, ddC had more adverse cellular effects than AZT. For example, at 10 PM ddC, there was an inhibitory effect on viability (Fig. 5A) and DNA content (Fig. 5B). Interestingly, this was not shown by an inhibition of the metabolic capacity measured by XTT-formazan production (Fig. 5C). Dextran sulfate, a known inhibitor of HIV infections in vitro, was evaluated under conditions identical to those of the nucleoside analogs. The results are shown in Fig. 6. Dextran sulfate protected CEM-SS cells against HIV-induced toxicity and infectious virus release as demonstrated by all parameters assayed (Fig. 6A-D).
97
A)
BCECF
DAPI
(Viability)
(DNA
Content)
140
z z x zD z .: x 2
100 60 60 40 7.0 120 I
cl
XTT 160 o : z n u 0 % z : .bp
(Metabolic
Activity) 400
-
140120-
/\
loo60 60 40 20
-
L&
/
-
o10-6
RT activity p24 production n infectious virus
$
300
”z
250
$ g
2oo 150
5 r
100 50
z 10-S
+ A
350
/
10-4
10 -3
0 i
1o-2
0
Uninfected
.
Infected
CEM-SS CEM-SS
Cells Cells
Fig. 7. Effect of PMA on uninfected (0) and HIV-infected (0) CEM-SS cells assessed after 6 days in culture using (A) cellular viability, (B) total DNA content, (C) metabolic activity (D) supematant reverse transcriptase activity (+), p24 antigen production (A) and synthesis of infectious virions (m). Values for infected cultures represent the mean of quadruplicate samples (SEM $8%), while uninfected values represent the mean of duplicate samples. All points are graphically represented as percentages of the uninfected, non-drug-treated control values.
The potential utility of multiparameter evaluations of anti-HIV compounds was most strikingly exemplified by results obtained with PMA (Fig. 7). Other studies have shown that PMA inhibits HIV-induced syncytium formation (Chowdhury et al., 1990) while enhancing virus production. Identical results were obtained in this study. While protecting CEM-SS cells from the cytopathic effects of HIV, PMA caused a marked decrease in DNA content (Fig. 7B) and cellular metabolic activity (XTT) of both uninfected and infected cultures (Fig. 7C). The BCECF viability assay revealed that this inhibitory effect was due primarily to a decreased rate of proliferation (cytostasis) rather than to overt cytotoxicity (Fig. 7A). Unlike the other four com-
98
pounds examined in this study, PMA appeared to have a biphasic response on virus production in the infected cultures. At lower concentrations, there was a marked increase in the production of infectious virus particles and viral antigens (Fig. 7D). At higher concentrations, however, virus production by infected cells was markedly inhibited.
Discussion The quest for new compounds, synthetic or naturally derived, with anti-HIV activity has led numerous investigators to search for rapid, reproducible, quantifiable and safe screening techniques. The ideal anti-HIV screening system would provide info~ation on multiple effects of test drugs on both the virus and host cell. Monitoring the direct effects of drugs on host test cells is important in that it allows for the normalization of detected antiviral effects with the number and viability of test cells. A variety of anti-HIV screening assays have been described which involve a single endpoint or a combination of endpoints. The metabolic reduction of the tetrazolium salt, MTT, has been used singly to evaluate anti-HIV compounds (Pauwels et al., 1988) or in combination with a supematant reverse transcriptase assay (Schwartz et al., 1988). Fluorescein diacetate (FDA) has also been used to evaluate cellular viability for both HIV-infected MT-4 cells (Schols et al., 1988) as well as human tumor cells (Larsson et al., 1989). As described in the introduction, the XTT-tetrazolium assay (Weislow et al., 1989) is currently used by the NC1 for the initial large-scale Stage I screening of synthetic chemicals and natural products for anti-HIV activity. The screen has been used to evaluate tens of thousands of compounds for anti-HIV activity and has effectively identified several new classes of antiviral compounds. The use of the XTT method offers advantages over the MTT procedure for large-scale anti-HIV screening from a biosafety standpoint as it requires fewer pipetting, aspiration or centrifugation steps. Past screening experience has shown that the XTT method, as with any colorimetric procedure, has a disadvantage when testing colored compounds such as the suramin analog dyes and many natural product extracts. Therefore, in our multiparameter battery of assays we employ a second, complementary assay of cellular viability to detect drug-treated HIV-infected cells using the carboxyfluorescein derivative BCECF. BCECF has recently been used to effectively determine the viability of drug-treated peripheral blood lymphocytes (Leeder et al., 1989) and was found to be more sensitive than the MTT assay. BCECF was chosen for use in our system for its linearity of fluorescence detection with very low numbers of cells (Fig. 3a) and also because its fluorescence was not quenched during fixation of HIVinfected cells with paraformaldehyde. Total cellular DNA content is measured using the A-T intercalating dye, DAPI. DAPI has previously been established as an accurate means to measure cell density of bovine and rat smooth muscle cells in vitro (McCaffrey et al., 1988). Averett ( 1989) developed an assay system, using the Screen Machine, to evaluate anti-HIV compounds using the nucleic acid dye propidium iodide (PI) to measure cytoprotec6on in combination with an antigen capture assay to measure HIV core p24 protein. DAPI
99
was chosen to measure total DNA content in this study because of its increased signal detection with small numbers of cells (Fig. 3) as compared with PI (Averett, 1989). In addition, cells stained with DAPI retain their fluorescence for long periods of time and can be stored for further analyses such as cell size comparisons. Certain classes of compounds being screened for anti-HIV activity may interfere directly with the endpoint analysis of any of the calorimetric, immunological, fluorescent and/orradioactive assays and therefore may give either a false positive or negative result. With this in mind it is clear that the best option is to obtain the most cellular and viral info~ation as possible to be able to compare and provide the broadest activity profile for candidate drugs. In our strategy for Stage II screening of new compounds we determine the relative effeets of known and potential antiviral compounds by the simultaneous assessment of cellular viability, metabolic activity and DNA content of virally infected, drug-treated lymphoblastoid cells, as well as measuring drug effects on infectious virus release, p24 antigen production and reverse transcriptase activity. Results obtained in the present study, using several known prototype inhibitors of HIV-induced cytopathology, establish the validity and suggest the general usefulness of these assays for initial follow-up confirmation and comparison of new compounds. AZT exerted a broad effect against acute HIV infections as indicated by the return to uninfe~tedcontrol cell values for viability, metabolic and DNAcontent assays (Fig. 4A-C) as well as the dramatic decrease of viral activity (Fig. 4D). The antiviral activity of ddC (Fig. 5) was also well documented in the multiparameter assay. However, at aconcentration of IO,UM, ddC caused adecrease in cellular viability (Fig. 5A) and DNA content (Fig. 5B) in both uninfected and infected cells. Similar results were seen with ddC-treated H9 cells chronically infected with HTLV-IIIB (data not shown). The ability of dextran sulfate to protect CEM-SS cells from the cytopathic effects of HIV in an in-well infection was demonstrated by all assay parameters (Fig. 6). Similar results were obtained when dextran sulfate was added to previously (bulk) infected CEM-SS cells (data not shown). These results are in agreement with a recent report which demonstrated that dextran sulfate can block both HIV adsorption and syncytium formation (Montefiore et al., 1990). The dete~ination of the direct effects of an antiviral agent on the target cells is important in the selection and overall evaluation of candidate compounds for drug development. For example, notwithstanding its tumor promoting properties, phorbol myristic acetate (PMA) has other properties which, if observed in isolation, might encourage its consideration as a potential anti-HIV agent. While PMA clearly protected GEM-SS cells from the cytopathic effects of HIV, there was also a dramatic direct effect of PMA on the cells themselves (Fig. 7B). As clearly apparent in the multiparameter analysis, the lower concentrations of PMA gave signs of protection as evident by an increase in viability (Fig. 7A) and metabolic activity (Fig. 7C). However, PMA appeared to stimulate the production and release of infectious virus and related viral proteins (Fig. 7D). In addition, PMA at moderate concentrations exerted strong growth inhibitor effects on both infected and uninfected cells. The assays described here are performed in conjunction with a Biomek 1000T~~ Workstation (Beckman), a semiautomated system for drug dilution, reagent addition
100
and plate-to-plate transfers and can be used for all steps including the addition of cells and virus solutions to test plates. In addition, the fhtorescence assays are performed using the Screen ~a~hineTM (Baxter Healthcare Corp.), a semiautomated fluorescent plate reader which has been described elsewhere (Averett, 1989; Leederet al., 1989). These semiautomated systems are ideal for work with HIV-infected cells to minimize direct technician involvement, to decrease the possibility of human error and more importantly to decrease possible exposure of staff to the lethal pathogen. The assays described provide a safe, semiautomated~ and reproducible muItip~ameter approach to screening compounds for antiviral activity. These assays provide a maximum amount of information with a minimum of sample, which is a critical issue when screening naturally-derived products or fractions thereof, where the amount of sample is often iimited. In addition, this multiparameter assay may aid in the prioritization of potential anti-HIV compounds for drug development. References Averett, D.R. (1989) Anti-HIV compound assessment by two novel high capacity assays. J. Virol. Methods 23,263-216. Boyd, M.R. f 1988) Strategies for the identification ofnew agents for the treatment of AIDS: a national program to facilitate the discovery and preclinical development of new drug candidates Forclinical evaluation. In: V.T. DeVita, S. Hellman and S.A. Rosenberg (Edsj. AIDS Etiology, Diagnosis, Treatment and Prevention, Lippincott, Philadelphia, pp. 305-3 17. Chowdhury, I.H., Koyanagi. Y., Kobayashi, S., Hamamoto, Y.. Yoshiyama, H, Yoshida, T. and Yamamoto, N. (1990) The phorbol ester TPA strongly inhibits HIV-l-induced syncytia formation but enhances virus production: possible involvement of protein kinase C pathway. Virology 176,126-l 32. Larsson, R. and Nygren, P. (1989) A rapid fluorometric method for semjautomated dete~ination ofcytotoxicity and cellular proliferation of human tumor cell lines in microculture. Anticancer Res. 9, 11I I-
i 120. Leeder, J.S., Dosch, H.M., Harper, P.A., Lam, P and Spielberg, S.P, (1989) Fluorescence-based viability assay for studies of reactive drug intermediates. Anal. B&hem. 177,3&l-372. McCaffrey, T.A., Agarwal, L.A. and Weksler, B.B. (1988) A rapid fluorometric DNA Assay for the measurement of cell density and proliferation in vitro. In Vitro Cellular Dev. Biol. 24,247-252. Montefiori, DC., Robinson, W.E., Modliszewski, A., Rowland, J.M., Schuffman, S.S. and Mitchell, W.M. (1990) Differential inhibition of HIV- I cell binding and HIV- I -induced syncytium fomlation by low molecular weight sulphated polysaccharides. J. Antimicrob. Chemother. 25.313-3 IX. Nara, P.L., Hatch, W.C., Dunlop, N.M., Robey, W.G., Arthur, L.O., Gonda, M.A. and Fischinger, P.J. (1987) Simple, rapid,quantitative, syncytium forming microassay forthe detection ofhuman immunodeficiency virus neutralizing antibody. AIDS Res. Hum. Retroviruses 3.2833302. Pauwels, R., Balzarini, J., Baba, M., Snoeck, R., Schols, D., Herdewijn, P., Desmyter, J. and De Clercq, E. (1988) Rapid and automated tetrazofium-based calorimetric assay for the detection of anti-HIV compounds. J. Virol. Methods 20.309-321. Rink, T.J., Tsien. R.Y. and Pozzan, T. (1982) Cytoplasmic pH and free Mg” in lymphocytes. J. Cell. Biol. 95.189-196. Schols, D., Pauwels, R., Vanla~~endonck, F., Balzarini, J. and De Clercy, E. ( 198X)A highly reliable, sensitive, flow cytometric/~uorometric assay for the evaluation of the anti-HIV activity of antiviral compounds in MT-4 cells. J. Immunol. Methods I 14,27-32. Schwartz, O., Henin, Y., Mare&al, V. and Montagnier, I,. (1988) A rapid and simple calorimetric test for the study of anti-HIV agents. AIDS Res. Hum. Retroviruses 14,441-44X. Weislow, OS., Riser, R., Fine, D.L., Bader, J.. Shoemaker, R.H. and Boyd, M.R. (19891 New soluhle-formazan assay for HIV-l cytopathic effects: application to high-flux screening of synthetic and natural products for AIDS-antiviral activity. J. Natl. Cancer Inst. 8 I, 577-586.
