Colorimetric assay fop cellular activity in microcapsules

Calorimetricassay for cellular activi~ in microcapsules Hasan Uludagand MichaelV. Sefton Daparrment of Chemical Engineering andApplied Chemistry, and Centre for Biometeriels, University of Toronto, Ontario M5S lA4, Canada (Received 4 May 1990; accepted 29 May 1990)

Cellular activity in microcapsules was determined by a simple calorimetric assay, based on the cellular transformation of a tetrazolium salt, 3-(4,5-dimethyl-thiazol-2-yl)-2,5-diphenyl-tetrazolium bromide, into an insoluble formazan which was quantified in a spectrophotometer. The results showed that when encapsulated Chinese hamster ovary fibroblasts were exposed to the tetrazolium salt containing tissue culture medium, the formazan crystals were formed inside the poly(hydroxyethyl methacrylate-methyl methacrylate) microcapsules; capsules containing no cells or dead cells formed no formazan. A detectable amount of formazan was readily obtained even from single capsules. Formazan production was dependent on the incubation time, but not on the amount of added reagent. Capsules from a high cell-density encapsulation (4 X 1 O6 cells/ml) formed more formazan than capsules from a low cell-density (4 X 1 O5 cells/ml) encapsulation, suggesting a positive correlation between the cell density and tetrazolium transformation in microcapsules. The tetrazolium assay indicated the maintenance of cellular activity but slow, if any, proliferation in microcapsules over a 2 wk testing period. Keywords: Microencapsulation,

membranes, cell function

Several procedures have been employed to microencapsulate a wide variety of cells and tissues’-5. The success of these procedures is measured by the ability of the encapsulated cells to maintain normal cellular activity once enclosed by a polymeric membrane. ln situ determination of cellular activity in individual microcapsules has, however, proved to be a challenging task. Conventional assays, such as tritiated thymidine incorporation, were considered inapplicable because of the small population of cells in microcapsules. This limitation was circumvented by the use of several indirect assays. Detection and quantification of secretory products was a useful index for cellular activity in microcapsules. Insulin and antibody secretion, for example, have been employed to evaluate the performance of encapsulated Islets of Langerhans and hybridoma cells, respectively’* 3. Urea formation, prothrombin and cholinesterase activity in external tissue culture medium have been determined to assess hepatocyte activity within microcapsules6. These procedures, involving radioimmunoassay or enzyme-linked immunosorbent assay (ELISA), are usually very sensitive but expensive and time-consuming. A minimum number of microcapsules (> 1 O-20 capsules) is required. Other techniques are invasive and involve breaking open the capsule wall to release the cellular contents7,8. The latter methods may damage the encapsulated cellular mass and require considerable skill, especially for strong, tough Correspondence to Professor M.V. S&on.

capsules. The development of a simple, non-invasive and rapid assay is of obvious importance for the evaluation of cellular activity in microcapsules. The 3-(4,5-dimethyl-thiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (M-IT) assay was originally used for intracellular localization of specific enzymesg. It was first shown to be a useful tool for quantitation of cell numbers by Mosman” and subsequently improved by several other authors’ ‘-13. It is based on the ability of mitochondrial dehydrogenase enzymes of living cells to convert MlT, a tetrazolium salt, into an insoluble formazan. This conversion requires an intact mitochondrial system and depends on the level of metabolic activity of the cells. The amount of formazan formed can be quantified and used as a measure of cell number and/or level of activation. The technique has been applied to a wide variety of cells, although its success has been sensitive to the operational parameterst4. An optimum working procedure needs to be developed for each type of cell under investigation. Here, the use of MlT assay for microencapsulated fibroblasts (Chinese hamster ovary, CHO) cells is reported.

MATERIALS

Biomaterials

1990, Vol 11 November

METHODS

Cell culture The stock culture of CHO cells (American Type Culture Collection, Rockville, MD, USA) was maintained on 25 cm2 0

708

AND

1990

Butterworth-Helnemann

Ltd. 0142-9612/90/090708-05

Assay

tissue

culture

USA)

flasks

(Corning

and fed with

a-MEM

Laboratories,

McLean,

fetal

serum

bovine

100

ng/ml

Glass

VA,

USA)

(Flow),

streptomycin

twice

a week.

incubating (Gibco)

The

3-4

transferred

were

monolayer

min.

The

with

Island,

medium

with

to flat-bottomed

and

NY, USA).

was

changed

weekly

0.25%

by

trypsin-EDTA

cells

96-well

10%

were

plates

(Corning)

with

reported4.

modifications. were

Briefly, pumped

ethyl

methacrylate-methyl

75%

HEMA,

MW

mg/l

NaCI; 1 150 The

bath

100

750

f

mg/l

assay

MlT

(Sigma,

in

microscopy,

Louis,

MlT

stock solution plates

and

as

from

distilled

water

reduce

USA)

filtration.

0.1

M

The

bath.

The

by

a

was the

crystals.

haemocytometer

Cell

before

analysis t-test and analysis of variance

to

evaluate

various

details

figure

statistical

populations.

(ANOVA)”

differences

The results

samples

are expressed

unless otherwise

of the experimental

was

(P < 0.05) as

indicated.

runs are indicated

in the

legends.

RESULTS Application

of MT’T assay to CHO cells in culture the applicability

series of experiments culture

insolubleformazan dependent

as

the

converted time-and

obtained

tissue

salt

into

concentration-

of cells, the amount

MlT

was added

incubation

were

MTT

time

into each

increased.

when

CHO

The

cells

were

with 25 ~1 of stock solution for 4 h. The production

of formazan

Figure

cells

as more

MTT

results

assay to CHO cells, a under standard

for a fixed number

increased

and

incubated

The

in an incubation

manner;

of formazan

of MlT

was performed

conditions.

as a function

of cell

1. Note that the formazan

with the number which

stirrer

deviation

of cells seeded from

linearity

absorbance

probably

of

because

is shown

at low cell densities, There

reading at zerocell

interference

of

with the absorbance

in

varied linearly

was observed.

small but significant components

number

absorbance

after was

a

number,

residual

medium

measurements.

by into a

overnight

with

was

prepared

as a

residues

The stock solution

MlT

at

was

3 wk of preparation.

and

in

37°C

100~1 for

conversion

aspiration

the wells

resultant

Georgetown,

in a loss of formazan

determined

mean f SD of triplicate

tissue

5 h unless

was stopped

of the

by washing

MTT

by

containing

the

capsules

in

for 10 min. This step was found to be critical

100 PI of dimethyl

M

mg/

to facilitate

undissolved

microcapsules

spectroscopic

The

were

step

plates; washing

the cells into microwells.

optimum

mg/l

determined

and incubated

incubated

by vacuum

medium

resulted

the washing

in 96-well

of 25 ~1 was added into each well of 96-

indicated.

the removal

except

for CHO cells grown

between

well

into the PBS.

were then transferred

MO,

containing

medium

otherwise

(0.1

10” were

surfactant

by a magnetic

in PBS. The

by sterile

well

In some

200 8000

p.p.m.

diameter

procedure

CHO monolayer adding

was obtained;

in all experiments,

medium.

St

removed

MlT.

x

flask containing CaCI,;

Ml, USA)

stored in the dark at 4°C and used within

to

The CHO

as a precipitation

Wyandotte,

stock solution

culture

glycol

at 4

MgSO,; 100

as

capsules/min)

from the hexadecane

50pm

culture

MTT

followed

of polymer of DMSO

nm within

spectroscopic

indicated.

same

To determine

capsules

(30

mg/l

GIBCO)

mm petri dish (Corning)

fresh tissue

were

The

kept in suspension

electron

5 mg/ml

at

twice for 30 min each with fresh PBS. The intact

capsules, scanning

problem

synthesized

ml volumetric

contained

of capsules

and washed

otherwise

No

S&on

poly(hydroxy-

medium

interface

Na2HP04;

were

to give

in polyethylene

culture

59.2

(Ll 01, BASF Chemicals, The capsules

1 05),

indicated.

KH,PO,;

mg/l

of

saline (PBS: 100

precipitation

the passage

due to the presence the blank was 100,ul

The

at 630

of DMSO.

M.V

for better capsule

Toronto Ont., Canada).

in tissue

otherwise

buffered

I KCI; 200

needle

methacrylate)(p(HEMA-MMA):

to fall into a 250

phosphate

unless

with reference

addition

and

The tip of the

of a precipitation

at a hexadecane/air

and allowed

a coaxial

provided

order

BDH Chemicals,

sheared

and

and dissolved

suspended unless

the cell suspension

solution was 10%

of the

previously’5

cells/ml

used

respectively.

and the elimination

(PEG-200;

was

inner needle was cut and polished

the needle tip. The polymer

cells were

and process

process

through

a conical shaped tip. This change

described

apparatus

The same

at 28 and 42 pI/min,

original chiba-type morphology

interference therefore,

employed

solution

assembly

of the

The Student’s

been

minor

polymer

min

Statistical

Microencapsulation have already

nm test wavelength

H. Uludag

or

used in microencapsulation.

The details of the microencapsulation

in microcapsules:

20-30

counts

either

activity

at 570

of 5% CO,

subcultured

trypsinized

(Flow

penicillin

in an environment

and the

cells

the CHO

for

unit/ml Grand

NY,

medium

supplemented

(Gibco,

air at 37°C

Corning,

culture

100

The CHO cells were incubated in humidified

Works,

tissue

for cellular

interference

formazan

sulphoxide

Ont., Canada)

experiments,

product (DMSO,

by vortexing

a specific

glycine

plus 0.1

NaOH)

was

M

added

from was

Caledon

of glycine

NaCl equilibrated to the

DMSO

in

while

crystals. The purple solution formed

Dynatech

ELISA

,

.

5

by

dissolving

was read on a

Alexandria,

VA, USA)

,

.

,

IO

well

1

under

Formazan normal

absorbance t,ssue

culture

CHO

cells

concentrations

of

overnight.

MTl

stock

incubated

for a further

DMSO.

The blank

solwon

.

,

15

as a function condit/ons. was (25

.

,

20

.

25

,

.

,

30

35

ofCells(x1000)

Number F/gun?

buffer

to pH 10.5

formazan

reader (No. MR600,

dissolved

.

Laboratories,

with a micropipette.

volume

0.0: 0

unconverted

added ul)

was

4 h. The formazan

ofnumberof Medium to

then

crystals

was CHO cells not exposed

cells seededper

(100

microwells added were

to MlTand

,uI) wth and

and

the

dissolved

vanow

mcubated cells

were

m 100 ,u/ of

dissolved

,n 100 PI

of DMSO.

Biomatenals

1990,

Vol

11 November

709

Assay for cellular activity in micro~a~o~es:

H. Uiudag and ML! Sefton

0.20 0.12

0.16 0.10

$ s f! g

0.12

9 0.06

0.04

0.00 j

T

1

T

3

2

4

Number of Capsules per Well

2.5

Figure 2 Formazan absorbance as a function of number of capsules per well. The capsules were either blank (m) or contained CHO cells (P) ~en~apsola~ion density: 4 X 10e cells/ml). Note that in the experimental range, the capsules containing CHD cells gave a higher formazan absorbance than blank capsules.

Application of MTT assay to microencapsulated CHO cells When microencapsulated CHO cells were exposed to MITcontaining tissue culture medium, formazan crystals were formed inside the capsules (light microscopy, not shown). Capsules containing no cells (blank capsules) or dead CHO cells, killed by overnight incubation with 0.5% KCN in saline, formed no formazan. The amount of formazan produced was quantified to determine the minimum number of capsules required to obtain a significant difference in formazan absorbance between the blank capsules and the capsules containing live CHO cells (Figure 2). The formazan absorbance with one to four capsules was significantly higher with encapsulated CHO cells than blank capsules. The difference between the blank signal and the signal from the cells increased as the capsule number increased. Since this difference was significant when even only one capsule was used, the remainder of the experiments were done with only one capsule per well. To optimize the assay parameters, the effects of MlT jncubation amount and time and the presence of glycine buffer in the extraction medium on formazan production and absorbance were investigated. There was no significant difference in formazan absorbance when different volumes of stock solution were added to microcapsules in 100 ,uI tissue culture medium (15-45~1, resulting in MlT concentration between 0.65 and 1.55 mg/ml), suggesting that

5.0

7.5

10.0

Incubation Time (hrs) Figure 3 Formazan absorbance as a function of M~i~cub~ion time. The formazan absorbance of blank capsules (m) was not dependent on the in~obatio~ time, whereas microencapsulated cells (@I gave increased absorbance at increased incubation times.

the formazan production by encapsulated CHO cells was independent of the amount of MIT present in the medium in the experimental range. However, the formazan production was strongly dependent on incubation time (Figure 3). Higher absorbance was obtained as the encapsulated cells were exposed to MlTfor longer periods. Note that there was no significant change in the formazan absorbance for the blank capsules as a function of time. The effect of glycine buffer on the formazan absorbance is shown in Table I. Addition of glycine buffer while dissolving microcapsules caused a significant (ANOVA, P < 0.05) increase in the formazan absorbance. The optimal volume of glycine buffer was found to be between 10 and 20~11. Nevertheless, the other results presented here were obtained without addition of glycine buffer. To determine the effect of cell number on formazan production in microcapsules, CHO cells were encapsulated at two different concentrations: 4 X 1 O5 and 4 X 1 O6 cells/ ml (Figure 4). The results are shown for three different encapsulation batches and are presented as the distribution of formazan absorbance, taking all batches together, among the individual microcapsules. It is clear that the amount of formazan produced is widely variable among the capsules. The mean + SD absorbances for low- and high-density capsules were 0.038 + 0.014 and 0.078 + 0.020 (n = 72, P < 0.001) respectively. There was no significant difference in formazan absorbance between the low-density capsules and blank capsules (0.041 + 0.012, n = 72).

Table f The effect of glycine bufferon the formazan absorbance of the micmencapsulated CHOcells. The microcapsules (one capsule per well) were incubated with M7Tandglycine was added while dissolving the formazan with DMSO. The wells to which no buffer was added served as control. The blank was 100 pi of 5MSO with a specific volume of buffer added. The addition of glycine buffer significantly increased the fofmazan absorbance by 20-40% Volume of glycine buffer added

Co&Jl 1Opl Formazan absorbance

710

Biomaterials

0.075

1990, Vol 11 November

i 0.009

0.105

i 0.013

2Opl

3Opl

0.110~0.012

0.096

it 0.010

Assay for cellular activity in microcapsules: H. Uludag and M.V Sefton

the difference

between

reached

significance

density

capsules

gave

blank or low-density period.

The

obtained

the blank and low-density

(P < 0.03) higher

formazan

capsules

highest

level

day 1) for high-density

high-

absorbance

than

at all times during the testing of formazan

absorbance

days 4 and 7 (P < 0.01

between

capsules

only on day 7. The

was

as compared

to

capsules.

DISCUSSION Microencapsulated 10

20

30

40

50

60

Formazan

70

80

90

Absorbance

100

110

120

MlT

130

x 1000

salt

into

production

Figure4 Distribution of formazan absorbance among the low- (4 X 1Os cells/ml) (m) and high-density (4 X 106 cells/ml) (S) microencapsulated cells. The absorbance was obtained from the individual microcapsules. Note that the capsules exhibited a wide range in absorbance at both high- and lowdensity encapsulation. However. there was a significant difference in formazan absorbance of the two populations. The data were obtained by selecting 20-25 capsules at random from three batches of low- and highdensity capsules after 24 h of encapsulation.

cells

in blank

suggested

activity

standard

dependent

O.llii!

0.14 -

f e

0.12-

5: 9 S

i

IA

, 2

1

.

,

.

6

4

,

.

,

.

,

10

8

.

12

,

.

,

14

16

Time (days)

total

of formazan utilized

Application proliferation followed

of Mll assay to assess in microcapsules

a high cell density low-density

which

unconverted was

amount

of

course to

assess

microcapsules. capsules

and

concentrations

of formazan

The MlT dependent

conversion

on

dependent

the

cellular

survival

Experiments capsules

for

(cells/ml)

and

were

carried

which

the

were

4 X lo5

2 wk

was

proliferation out using

in

blank

The

latter

was not limited by the

but

rate

that MTT (MW

and

the addition resulted capsules.

for blank capsules, (ANOVA,

drop of absorbance

whereas

P > 0.05)

density

microcapsules

capsules

resulted

was

observed

there was no significant

in the formazan with

time.

in higher formazan

absorbance Although absorbance

change

after day 2,

the

low-density

capsules MlT,

was

parameters might

that

the

shown

range

formazan It was

the capsule was present

might

cells. For example,

about the effect

with

improved

cell by

solution which from

the previously

of glycine

longer

metabolizing

be further

absorbance

in accord

and time might

this assay was applied

of glycine buffer in the extraction formazan

Since

to vary significantly

volume

when

not

strongly

of the substrate

through

be used for slow

The sensitivity

This was

was

but

of CHO cells in microcapsules.

cells, the incubation

in higher

observation

buffer

individual reported

on formazan

absorbance”. In the development

for low-

low-density

culture

transformation.

diffusion

types of encapsulated

low- and high-density

(P < 0.005)

414)

times

Figure 5 shows the time course for

significant

from

by the availability

incubation

(low density)

A

1 O5

absorbance

time in the experimental suggested

to different phenotypes.

of

of cellular

cleavage

important

cell

capsules.

tissue an

fast, so that excess substrate

number

different

become

4 x 1 O6 (high density). of blank,

X

to overshadow

by individual

results

MlT,

three

batches

gave

enough

concentration

production

solution

core

capsules (-4

assay for CHO cells.

produced

on the incubation

studied.

the rate of MlT

for

significant formazan

blank

density

of residual

MlT

of CHO detection

capsules.

among

production

It was

by microcapsules

and

of the MlT

was

cellular

microcapsules

in formazan

capsules

that a small volume

and/or

the

a low concentration

a lower limit of encapsulation

medium

cellular

concentrations.

difference

for the applicability

reading

between

was produced

than

and

This sensitivity

in individual

cells at different

of

a significant

to quantify

a relation

production

the amount

to detect

therefore,

that more formazan

for the small

cellular

was

to MTT

in microcapsules

microcapsule.

to obtain

wall was relatively

time

activity

and,

further

the

increased,

It was possible

even in a single

believed

The

cellular

activity,

small

Figure 5 Formazan absorbance as a function of time from blank, low(4 x 1O5 cells/ml) and high- (4 x 1O6 cells/ml) density capsules. The highdensity capsules gave highest formazan absorbance at all times during the testing period. Only on day 7 did the difference in formazan absorbance between blank capsules and low-density capsules reach a significant level. The data represent mean + SD of individual microcapsules from three different batches and were generated by randomly choosing 20-25 capsules at each time pomt. (+) Blank, (0) 4 X 1O5 cells/ml, (0) 4 X 1OGcells/ml.

exposed

production.

It is possible

.

In the case of

production

the

cells/ml)

.

cell model.

between

suggested

o.oo! 0

in

the applicability

formazan

number

and

grown

formazan

between

0.02-

indicating

of capsules

cells. The lack of significant

0.04-

cells

as the capsule

with

.

CHO

formed also increased, which indicated a correlation

shown 0.06

by the

formazan

by encapsulating

r;l

conversion

the cell number

for

the

number

activity and formazan

0.08-

between

obtained

cells,

on the

dead cellular

solution;

amount

0.10 -

of passive

assay for the chosen CHO

containing required

of the tissue culture medium.

tissue culture conditions,

of the MlT

soluble

lack of formazan

or capsules

correlation was

The

transformation

not a result

conversion

encapsulated 0.18 -

the

or other components

A positive MlT

formazan.

capsules

that

and was

polymer

were able to transform

CHO cells

insoluble

of the present

were able to obtain valuable ation

procedure

information

and subsequent

Biomatenals

assay, the authors about the encapsul-

cell behaviour

in micro-

1990, Vol 11 November

711

Assay for cellular actti@ in mjc~~apsules:

H. Uludag and M.K Sefton

capsules. A large variation in the formaran production among individual capsules within each batch was detected. The preliminary experiments suggested that this variation was closely related to the efficiency of polymer solution to entrap the cell mass in aqueous core. It appeared to be possible to minimize, but not to eliminate, the variation among capsules by increasing the encapsulation efficiency through adjustment of the cell:polymer flow-rate ratio. The MlTassay also indicated the maintenance of cellular activity in microcapsules throughout the 2 wk test period. It was noted that the extent of M~conversion increased with time and reached a plateau level between 4 and 7 d after CHO cells were encapsulated at relatively high density (4 X IO” cells/ml). These changes in formazan production can be attributed either to cell proliferation or to activation of the cell metabolism. However, the fact that capsules from the lowdensity encapsulation did not show a significant increase in formazan production even after 14 d suggested that the rate of proliferation was very low. Attention has now been directed to understanding the mechanism of these changes in formazan production. Care should be taken in using formazan absorbance as a measure of cell number in microcapsules. To use a calibration curve derived under normal tissue culture conditions, it is necessary to assume that each cell retains the same level of metabolic activity once enclosed by a polymeric membrane. Upon encapsufation, if the cells are activated or suppressed, the correlation of formazan absorbance of capsules with the calibration curve will result in either over- or underestimation of the cell numbers, respectively. In addition, the time delay for the diffusion of MlT through the capsule wall has to be taken into account. In the current work, the CHO cells survived in suspension in microcapsules, unlike normal tissue culture conditions in which the cells proliferated in an attachment-dependent manner. Since such a drastic change obviously affected the cellular metabolic activity, the authors did not attempt to calculate microencapsulat~ cell numbers from the calibration curve.

ACKNOWLEDGEMENTS

CONCLUSION

12

A calorimetric assay, commonly referred to as the MIT assay, was modified and used to assess cellular metabolic activity in microcapsules. Based on the cells’ ability to transform a tetrazolium salt into formazan, the assay was shown to provide quantitative information about the encapsulated cells. The ability to detect metabolic activity even from individuat capsules offers an obvious advantage over techniques previously utilized. However, caution is necessary in interpreting the results of the assay, since the metabolic state of the cells affected the amount of formazan production per cell. The simplicity and low cost of the present assay makes it a valuable tool to screen a large number of encapsulation parameters in a short period of time.

13

712

Biomaterials

1990, Vol 11 November

The authors acknowledge the financial support of the Ontario Center for Material Research, National Institute of Health (AM 29689) and Medical Research Council of Canada.

REFERENCES I 2

3

4

5

6

7

8

9 10

17

14

15

16 17

Lim, F. and Sun, A., Microencapsulat~ islets as bj~~ificial endocrine pancreas, Science 1980, 210, 908-910 Stevenson, W.T.K., Evangelista, R.A., Sugamori, M.E. and Sefton, M.V., Microencapsulation of mammalian cells in a hydroxyethyl methacrylate-methyl methacrylate copolymer: Preliminary develop ment. Biomater. AttX Cells Artif Organs 1988, 16(4), 747-769 Yoshioka, T., Hirano, R., Shioya. T. and Kaka, M., Encapsulation of mammalian cells with Chitosan-CMC capsule, &ore&. Bioeng. 1990, 35.66-72 Crooks, C.A., Dougles, J.A., Broughton. R.L. and Sefton, M.V., A submerged jet processforthemicroencapsulationof mammaliancells in a HEMA-MMA copolymer: Effect of capsule morphology and permeability, J. Biomed. Mater. &es. 1330, 24. 1241- 1262 Dupuy, B., Gin, H., Baquey, C. and Ducassou, D., In situ polymerization of a microencapsulating medium round living cells, J. Biomed. Mater. Res. 1988.22, 1061-1070 Cai, 2.. Shi, Z., Sherman, M. and Sun, A.M., Development and evaluation of a system of microencapsulation of primary rat hepatocytes, Hepatology 1989, 1 O(5), 855-860 Dawson, R.M., Broughton, R.L., Stevenson, W.T.K. and Sefton, M.V., Microencapsulation of CHO cells in a hydroxyethyl methacrylatemethylmethacrylate copolymer, Biomaterials 1987, 8, 360-366 Gorelik, E. et a/, Microencapsulated tumour assay; New short term assay for in vivo evaluation of the effects of anticancer drugs on human tumour cell lines, Cancer Res. 1987,47(21), 5739-5747 Altman, F.P., Tetrazolium salts and formazans, Prog. Histochem. Cytochem. 1976, S(3). 1-51 Mosmann, T., Rapid cdorimetric assay for cellular growth and survival: Application to proliferation and cytotoxic assays, ./. /mrnuRo~ Methods 1983, 66, 55-63 Danizot, F. and Lang, R., Rapid calorimetric assay for cell growth and survival: Modifications to the tetrazolium dye procedure giving improved sensitivity and reliability, J. immunol. Methods 1986,88, 271-277 Twentyman, P.R. and Luscombe, M., A study of some variables in tetrazolium dye (MTT) based assay for cell growth and chemosensitivity, Br. J. Cancer 1987, 56, 279-285 Carmichael, J., DeGraff, W.G., Gazdar. A.F., Minna, J.D. and Mitchell, J.B., Evaluation of a tetrazolium-based semiautomated calorimetric assay: Assessment of chemosensitivity testing, Cancer Res. 1987, 47.936-942 Hansen, M.B., Nielsen, SE. and Berg, K., Re-examination and further development of a precise and rapid dye method for measuring celf growth/celt kill, J. lmrn~no~ Methods 1383, 119, 203-Z 10 Stevenson. W.T.K., Evangelista, R.A., Sugamori, M.E. and Sefton, M.V., Preparation and characterization of thermaplastic polymers from hydroxyalkyl methacrylates, J. App/. PO&m. Sci 1987, 34, 65-83 Croxton. F.E. and Cowden, D.J., Applied General Statistics, PrenticeHall, New York, USA, 1939, p 35 1 Plumb, J.A., Milroy, R.. and Kaye, S.B., Effects of the pH dependence of 3-(4,5-dim~hyl-thiazol-2-yl)-2.5-diphenyl-tetrazolium bromideformazan absorbtion on chemosensitivity determined by a novel tetrazolium-based assay, Cancer Res. 1989, 49, 4435-4440