-------
S3N3ZN380dCnHOia
3N3HAH130dO~IHDVdl31
3N3mOJL
3au~msia
24
-------
^-^
CO Q
OS W
H EH
EH U
S3
fc,
O -
CO
en OS
S3 W
0 EH
H <
EH &
$ EH
EH CO
* g!
U
S Q
O 2
U <
in
w
PQ
^
EH
1
O CO 1
M Q) CU
0 C -
rH Q) 1
rCj N £
0 C -
•H Q) 1
Q X! O
* 1
to QJ
d) •»
C 1
0) g
H •>
>-, |
X O
^f
1 — 1
u
CN
U
0)
CD
3
O
EH
g
O
4H
O
!H
O
•^
U
QJ
rH
p.
g
-4H
QJ OJ O
1 i Ij j 1 j
4J 0 4J g
($ fd ($
H g QJ rH OJ
ft (d tn fc >-i
QJ (d 4-J
X! M S X! W
4-1 4-> QJ 4-> C
d w co d S
o a o o
CO ^ CO 13
VD
^.
oo
OJ
•*
m
m
o
*
o
*.
H
•
O
».
in
o
•
0
o
•5J*
rH
(M
•
O
QJ
Cn
td
S -P
QJ C
QJ CO 01
CT> d
fd ^4 rH
£ QJ 4H
Q) > MH
CO C QJ
QJ
Q
^
*.
00
k.
ro
rH
•
o
V
CM
•
O
k»
CN
•
0
CN
•
O
CN
•
O
(N
1 tl
rH QJ
QJ g >
^ td G
QJ QJ
G r( Q
O 4-1
-P W E
CCO
!H £ SH
O O MH
EH v**'
25
-------
A
nr
o
6O
12O
TIME
Figure 10. Volatile organics in water from
(a) Thornton well, 3-1, (b)South
Platte River near well, (c) Denver
sewage plant effluent. Peaks iden-
tified: (A) chloroform (B) toluene
(C) tetrachloroethylene (D) m-and
p-xylene (E) ethyl benzene (F) o-
xylene (G) dichlorobenzene.
26
-------
Water from an experimental reverse osmosis unit at the
Denver sewage plant showed marked peaks at positions correspond-
ing to chloroform, toluene and tetrachloroethylene, as well as
lesser peaks, indicating that reverse osmosis did not remove
the more volatile compounds. Activated carbon treatment does not
remove these compounds either. Liquid chromatography, however,
showed that carbon filtration and reverse osmosis were effective
in removing the less volatile and more polar impurities from
water (see below).
Tap waters from Denver and Boulder were examined and showed
very little volatile content, well below 1 ug/£. The city of
Boulder, Colorado, has an unusually clean source of drinking
water. Humic material is present, but there is little contami-
nation of human origin. The water is chlorinated and clarified
with alum before distribution. No toluene or other aromatic
hydrocarbons were found at any stage of the treatment, though
trihalomethanes were found in low concentrations. It seems that
precursors of toluene may exist in wastewater but not in un-
contaminated natural waters.
Pristine mountain water from a spring in the timber at
8,000 feet elevation (2,400 meters) showed five peaks, two of
high volatility emerging near the beginning of the temperature
program, and three of low volatility emerging near the end. The
concentrations were about 0.1 yg/fc. Identifications were not
attempted, but one may speculate that the compounds came from
the decay of humic material. Similar peaks of low volatility
appeared in Denver tap water. Water from Grand Lake, Colorado
was almost devoid of volatile organic matter.
NON-VOLATILE COMPOUNDS
Non-Volatiles by "Trace Enrichment"
The trace enrichment technique depends on the absorptive
properties of porous silica coated with chemically-bonded
octadecyl groups (C-, „) . This material is a universal absorbent
for all organic compounds except those that are very water-
soluble, ionized or highly polar, and it is widely used as a
column packing in liquid chromatography. If water that contains
traces of dissolved organic compounds is pumped through a tube
or column packed with octadecyl silica, the less polar compounds
are retained. Even phenol and caffeine, compounds that dissolve
significantly in water and are considered to be fairly polar,
are held quite strongly. One can pump a large volume of a very
dilute solution, of concentration 1 ppm or less, and find that
these compounds are quantitatively retained.
The absorbed compounds can be released and stripped from
the column by passing a water-miscible organic solvent like
methanol or acetonitrile. Mixtures of the solvents with water
27
-------
may be used, and a selective release of the absorbed substances
is accomplished by adding methanol to water in progressively
higher proportions, starting with pure water and ending with pure
methanol, according to a solvent program or gradient. The most
water-soluble or polar compounds leave the column first, and the
most hydrophobic or non-polar compounds leave last. Each com-
pound or group of compounds can be recovered as a relatively
concentrated solution, and the concentration can be further
raised by evaporating the solvent. Organic compounds originally
present in, say, five liters of water can be concentrated into
volumes of 2 m£ or less.
The apparatus we use is shown in Figure 11. The pumps are
Waters Assoicates Model 6000 A adjustable-speed liquid chromato-
graphy pumps; they are controlled by a Waters Associates Model
600 Solvent Programmer. The column is of stainless steel, 50 cm
long, 1 cm internal diameter. Following the column is an ultra-
violet liquid chromatography detector; normally this was a
single-wavelength detector reading the absorbance at 254 nm,
though a Schoeffel variable-wavelength detector was sometimes
used. The signal from the detector was recorded. A typical
record obtained with wastewater and most city water supplies is
shown in Figure 12. The significance of the record is discussed
in the next section.
The column was packed with Bondapak C^g-Porasil B, from
Waters Associates. This material has a large particle size, 37-
75 pm, and is usually used for preparative chromatography. Its
chromatographic resolving power is rather poor; it is rated by
Waters Associates at 350 plates per meter. We chose it, rather
than the' high-resolution Micro-Bondapak C-.Q, because it absorbed
and desorbed the constituents of wastewater reversibly and never
clogged up. In our first summer's work we used expensive, pre-
packed, microparticulate, analytical columns and ruined two of
them; an irreversible sorption took place after pumping several
liters of water, the back pressure rose to an intolerable level,
and we were not able to clean and unplug the columns. We have
used the Bondapak C,g-Porasil B column for two years and it
flows as easily as ever; the back pressure with water at 7 m£/min
is about 600 psi. The brown humic materials of wastewater are
reversibly absorbed and desorbed. After a year's operation we
flushed the column with acetonitrile, chloroform and hexane, and
extracted 100 mg of white, waxy solid whose infrared spectrum
was that of a straight-chain paraffin with some -C-0- linkages.
This was the only indication of irreversible absorption.
The Bondapak was packed dry into the column; no special
equipment or technique was needed. Though the resolving power
is modest, it is sufficient for our present needs. We tested
other packings, including graphitized carbon and the porous
polystyrene gel Hitachi-3010, and found them to retain humic
materials irreversibly.
28
-------
MIXER
PUMP
A
SAMPLE
WATER
PUMP
B
METHANOL
SOLVENT
PROGRAM
1
COLUMN
SILICA-C,8
37-75 MICRONS
RECORDER
Figure 11. Apparatus for trace enrichment.
29
-------
SALTS
OF ACIDS
(R COO No)
IONIC AND
HIGHLY-
POLAR
COMPOUNDS
HUMIC ACIDS
POLAR COMPOUNDS
WEAKLY-POLAR
NON-
POLAR
COMPOUNDS
SAMPLE !H2°! GRAD.ENT I
METHANOL
(TO 2 LITERS) ' Q% (30 min.) ^ |00o/o
CH3 OH
Figure 12. Chromatogram obtained in trace enrichment.
30
-------
We went to the large-sized column (1 cm x 50 cm) to produce
fractions that would be large enough for toxicity tests and for
more detailed chemical examination. Our practice is to collect
a relatively small number of fractions on the first operation,
to test each fraction for toxicity and mutagenic action, and
then to subdivide each fraction further, paying special atten-
tion to the more toxic fractions. The mode of operating the
column is summarized here:
Sample Preparation--
Filter under suction through glass fiber filters, finishing
with 0.7-micron filter. Before the last filtration, heat the
sample to 70° - 80°; suction filtration then removes dissolved
air. Finally, adjust the pH to 7.5 - 7.8 by adding nitric acid.
Filtration is necessary to avoid blocking the pumps or the
column; dissolved air must be removed to avoid vapor locks in
the pumps. The pH adjustment is needed to avoid chemical attack
on the porous silica packing (which occurs above pH 8) and to
convert the weak organic acids of the sample to their anionic
forms. The effect of pH on the elution pattern will be de-
scribed below.
In some experiments where we wanted to process a large
amount of sample, we performed a preliminary concentration by
fractional freezing, using a Model 3-1000 freeze concentrator
(Virtis Co., Gardiner, NY). This operation was done at the
Water Quality Laboratory of the U.S. Geological Survey in
Lakewood, Colorado. The "freeze concentrate" was one-tenth the
volume of the original wastewater and was difficult to filter
because of the high concentration of detergents and suspended
clay. Most of the suspended matter was removed by centrifuga-
tion, after which the filtering was relatively easy.
Sample Loading—
Pump the desired volume of filtered sample through the
Bondapak column at 7 m£/min., using one Waters Model 6000 A
pump. In the early work the volume was 2 liters of wastewater
or 1 liter of 10:1 freeze concentrate. Later we routinely
pumped 6-10 liters of filtered wastewater to get enough material
for further study.
Another way to handle larger samples is to use an auxiliary
"collector column" 15 cm x 1.0 cm internal diameter, packed with
the same Bondapak-Cig that was used in the main analytical
column. We pumped sewage effluent through this column without
deaerating it or adjusting the pH; however, between the pump and
the collector column were three tubes packed with filtering
materials, glass beads, sand, and glass wool. The pump was a
Madden diaphragm pump that was insensitive to suspended matter
and could be left to run all day without attention. After pump-
ing 3 gallons (12 liters) of secondary sewage effluent through
31
-------
the collector column at 0.5 liter per hour we disconnected the
column, flushed it with distilled water, connected it to the in-
let of the analytical column and started the water-to-methanol
program that is described in the next section.
A collector column of Bondapak-C,g was used by May, et al.
(5) to concentrate poorly-volatile hydrocarbons from sea water
and marine sediments; they attached the loaded column to an
analytical column of MicroBondapak C, g and applied a water-to-
methanol gradient, as we have done. use of a separate collector
column is clearly more convenient than pumping the whole sample
through the main column, but with wastewater, the suspended
matter causes a serious problem. The "on-line" filters that we
have used so far are inefficient. Our collector column became
fouled with black material after a few gallons of wastewater had
passed, and it lost its effectiveness. We shall experiment with
better on-line filters to see if this fouling can be avoided.
Flushing, Gradient Elution—
With the solvent programmer activated and the methanol
pump (B) (see Figure 11) ready to operate, reduce the flow rate
to 5 mJl/min and turn the solvent selector valve of the water
pump (A) to admit pure water. (This was a house distilled water
that had been passed through a bed of Amberlite XAD-2 resin and
redistilled in glass.) Pump pure water for 15 minutes, then
start the water-to-methanol gradient, a linear gradient from 0%
to 100% methanol in 30 minutes. Meanwhile, run the UV detector
and recorder, and collect samples as appropriate. We use the
scheme shown in Figure 13.
Testing the Fractions—
Collect each fraction in a glass-stoppered bottle, and
concentrate to the final volume desired (usually 2 m£) on a
rotary evaporator (Buchi Rotavapor-M). Transfer the concentrat-
ed fractions to 4-milliliter glass vials for storage, in treat-
ing the last fractions (G5 in Figure 13), where the solvent is
practically pure methanol, see if solids that stick to the glass
evaporation flask separate and are not transferred to the
sample vial. If this is happening, rinse the evaporation flask
with small amounts of 95% ethanol to ensure that all solutes are
removed and transferred. Note the approximate alcohol concen-
tration for the information of those who will perform toxicity
tests.
It is in the form of these 2-milliliter concentrates that
the fractions are submitted for mutagenicity and toxicity test-
ing. For chemical testing (secondary chromatography) we pre-
ferred to evaporate fractions G4 and G5 nearly to dryness and
take up the residues in methanol.
32
-------
FLUSH
Figure 13.
GRADIENT
Trace enrichment chromatography
(schematic) with system of
fractions taken.
33
-------
The Liquid Chromatography Elution Record
The chart-paper record of UV absorbance against time and
solvent volume looks like Figure 12 (schematic) or Figure 14
(an actual wastewater). If the column is originally filled with
pure water, and wastewater is then introduced, the UV absorbance
rises sharply, as soon as the wastewater enters the detector,.
and eventually levels off. We see that some UV-active material
passes through the column without being absorbed at all, while
other material is held weakly by the column. The column is
soon saturated with these weakly-absorbed substances.
The substances that pass through the column during loading
account for some 60-75% of the total organic carbon in the
sample. Total organic carbon was measured with a Dohrmarm
Envirotech Analyzer, Model DC 52-D. They do not appear to be
toxic, and we have therefore not tried to characterize them, but
we can expect them to be salts of relatively strong acids (like
uric acid, pK = 3.9) and substances that are very soluble in
water, like sugars, which were found in wastewater by Pitt and
others (11). Sugars and many other organic compounds would not
be "seen" in the ultraviolet. Inorganic salts likewise pass
through the column without being retained or "seen" in the ultra-
violet.
After the sample has been pumped we pass redistilled water
for 15 minutes to flush the column before beginning the methanol
gradient. As soon as the distilled water reaches the detector
the UV absorbance rises abruptly to a high maximum, then falls
more gradually. We call this peak of absorbance the "flush
peak". We have found the following facts about the material in
the "flush peak":
- It is moderately toxic
Its pH is about 8
Its UV spectrum is changed by adding acid, suggesting
that salts of weak acids are present.
The total organic carbon accounts for about 5% of the
organic carbon originally in the wastewater.
The peak height is decreased if the pH of the sample is
lowered.
Several substances are present, and they can be partial-
ly resolved by liquid chromatography on a column of
anion-exchange resin. Some of the peaks from the resin
column can be shown to be due to anions, others to
neutral compounds. A refractive index detector placed
in series with the ultraviolet detector reveals no new
peaks; that is, the compounds of the flush peak are all
aromatic in character.
The solution flushed from wastewater is light brown in
color.
34
-------
x|.28
FLUSH GRADIENT
i
Figure 14.
I
Liquid chromatogram of one
liter of Pomona California
secondary effluent; pH 7.8.
35
-------
We were able to produce an artifical "flush peak" from dis-
tilled water by adding benzoic acid, but only if we also added a
salt like sodium sulfate or nitrate in a concentration 0.001 M
or so. The UV spectrum of this flush peak was that of the
benzoate ion, not benzoic acid.
Clearly, the "flush peak" represents material that is weak-
ly absorbed by the Bondapak-C^g and is released by passing pure
water. The role of electrolytes in forcing the retention of
benzoic acid, which is subsequently released by flushing with
water, is not obvious and may involve the absorption of sodium
benzoate ion pairs. If we flush with a salt solution instead of
pure water, no peak appears.
When the methanol gradient is started the UV absorbance
first drops, then rises, at first rather slowly, then rapidly.
The color of the effluent is at first very pale, then, after the
UV absorbance has been rising for 2 minutes, it becomes deep
brown. Most of the brown material in the wastewater is concen-
trated in this peak, which we label "Gl". (The G stands for
"gradient".) After the UV absorbance reaches its maximum and
begins to fall, generally 15 minutes from the start of the
gradient, the effluent color changes sharply from brown to blue-
green. It remains green for 2-3 minutes (10-15 m£), then
changes to light brown. Sometimes the absorbance rises to give
a small "hump" while the green fraction is passing. We collect
the green fraction separately. It is toxic and possibly muta-
genic. It has appeared in every wastewater sample we have test-
ed, and its visible absorption spectrum shows a sharp peak at
630 nm. Gel permeation chromatography (see section entitled
"Secondary Chromatography and Peak Identification") shows it to
contain at least two substances, a green component of low
molecular weight and a pale brown component of high molecular
weight.
Following the green fraction, the UV absorbance falls and
eventually drops to zero some 10 minutes (50 m£) after the
gradient has ended. (Between the time the liquid leaves the
pump and the time it reaches the UV detector, 30 m£ have passed
and 6 minutes have elapsed. Most of the absorbed organic mate-
rial has been stripped from the column, therefore, before the
methanol concentration reaches 100%).
The last fraction, G5 of Figure 13, contains the substances
that are least polar, and it is here that most toxicity is
found. The G5 fraction is a very pale pink, almost colorless.
All the other fractions have a brown color (except green G2),
showing that the humic material spreads itself all over the
gradient. It complicates the chemical analysis, and frustrates
attempts to do gas chromatography-mass spectrometry.
The general form of the elution record for wastewaters is a
36
-------
broad wedge-shaped peak on which is superimposed smaller peaks
and minor irregularities. The "envelope" of the peak may be
very largely due to humic substances, but we recall that the
resolving power of the column is not great, and many individual
peaks will merge into a continuum. At this point, the width and
form of the "humic acid" peak depends on the pH of the sample
that is loaded. The lower the pH, the wider and higher the
peak. Figure 15 shows the effect. At the lowest pH (3.9) the
envelope is broad and carries at least five distinct peaks. The
flush peak, meanwhile, is small. As the pH is raised, the peaks
in the gradient region disappear one by one, and the envelope is
thinned down. At the same time the flush peak grows higher.
The obvious interpretation is that weak acids are present that
are retained by the column in their undissociated forms (HA) but
not in the ionic forms (A~). The ionic forms come out in the
pre-flush stage or are somehow transferred to the flush peak,
perhaps through the retention and subsequent release of ion
pairs.
The effect seen in Figure 15 makes it necessary to control
the pH of the samples before they are- loaded on to the column.
A high pH is preferred to a low pH, because toxic substances are
likely to be non-ionic compounds like chlorinated pesticides and
polycyclic aromatics, whose retention is independent of pH, and
at high pH there is minimum retention of unwanted humic acids.
The pH may not be raised above 8, we repeat, because the bonded
silica column packing would be attacked.
Elution of Known Compounds
A number of known compounds were loaded onto the column by
the standard procedure, starting with distilled water or tap
water "spiked" with the compounds at concentrations of 1-2 ppm.
The elution times depended on the size of the column, the flow
rate, and the gradient. Times were measured early in this
research, and different columns and conditions were used. The
sequence of elution was the same in all the columns, however.
Table 6 shows the elution sequence, and approximate retention
times referred to our standard 1-cm by 50-cm column.
Compounds that eluted early, like phenol and caffeine, may
have moved down the column during the water flush, but doubling
the flush period reduced the retention time of caffeine by less
than one minute. The prescribed 15 minute flush period is,
therefore, not critical.
The sensitivity of detection was estimated by introducing
known 5-microgram quantities of caffeine. The peak height under
standard operating conditions was 0.02 absorbance units at
275 nm. In two liters of water this represents a concentration
of 2.5 ppb (yg/£). This is very satisfactory, but we should
note that in high performance liquid chromatography we detect
37
-------
pH 7.9
GRAD.
GRAD.
GRAD.
Figure 15. Liquid chromatogram of Denver secondary sewage
effluent showing the effect of pH on retention;
full-scale absorbance - 0.64 unit.
38
-------
10 ng of caffeine, because the peaks are so much narrower with
short columns and fine packings.
Liquid Chromatography and Tqxicity of Treated Wastewaters
In October, 1977, the Denver sewage plant tested, on a
small scale, the effects of reverse osmosis and activated carbon
on the secondary effluent. We obtained samples of the products
and ran chromatograms, with the results shown in Figure 16.
Both treatments reduced the organic content considerably. Com-
paring Figure 16 with Figure 15, for instance, one should note
TABLE 6. ELUTION SEQUENCE ON C-18 COLUMNS
Compound
Minutes from start
of gradient
Phenol
Caffeine
Benzyl alcohol
Benzaldehyde
Meta-cresol
Meta-chlorophenol
Ethyl benzoate
Diethyl phthalate
Atrazine
Naphthalene
Dibutyl phthalate
Diphenyl oxazole
Pyrene
19
20
22
25
26
28
29
30
30
32
34
35
36
Note: This is a composite of results from three columns packed
with octadecyl silica. The elution times are scaled to our
"standard" 1-cm x 50-cm column at 5 m£/min with a 30-minute
linear gradient.
39
-------
TREATED
SEWAGE
EFFLUENT
xO.08
ABSORBANCE
FULL SCALE
CARBON-
TREATED
(1500ml)
REVERSE
OSMOSIS
(1100ml)
Figure 16.
GRADIENT
Chromatograms of Denver secondary sewage
effluent after the additional treatments
shown.
40
-------
that full-scale deflection was only 0.08 absorbance units in
Figure 16, but 0.64 units in Figure 15.
Reverse osmosis suppressed the flush peak and the pre-flush
absorbance, indicating that it removed all electrolytes, in-
cluding inorganic salts and ionized organic acids. The salt
'content was checked by titration and was below 10 N. Carbon
treatment does not remove inorganic salts and only partially re-
duces the inorganic acid content (as indicated by the flush
peak). Both treatments remove the big humic acid peak, and both
let pass some material later in the gradient, corresponding to
our G3 fraction. We were not able to make toxicity tests with
these fractions because the Denver pilot plant shut down just
after we took our samples.
Samples were also obtained, before and after carbon treat-
ment, from the Pomona, California treatment plant. The liquid
chromatography curves showed carbon treatment to be effective
(see Figure 17). The treatment does not remove the flush peak,
though it thins it down, and it does not remove the most polar
part of the humic material, but it does remove the less polar
material, especially that corresponding to our G4 and G5 frac-
tions. Measurements of total organic carbon showed an overall
reduction of about one-third, the pre-flush or polar material
being reduced only 20%.
The tests of toxicity showed (1) that most of the toxicity
was at the non-polar end of the gradient, and (2) that carbon
treatment reduced this toxicity.
To test reverse osmosis we purchased a laboratory water
purifier, Milli-R04, from the Millipore Corporation, Bedford,
Massachusetts, and modified it so that a limited volume of feed
water could be re-circulated. Eight liters of Denver secondary
sewage effluent were treated and divided into 4 liters that had
gone through the membrane (the permeate) and 4 liters that had
not gone through the membrane (the reject). The chromatograms
are shown in Figure 18. The volumes indicated are those of the
samples that were loaded onto the chromatographic column; thus,
the difference between the permeate and the reject is greater
than the chromatograms show.
Reverse osmosis reduced the total organic carbon consider-
ably, and removed electrolytes, suppressing the pre-flush absor-
bance and the flush peak. Once again, reverse osmosis allowed
some material to pass in the middle of the gradient, and there
was evidence that non-polar and weakly polar substances were
passing. Toxicity tests confirmed that some of the undesirable
weakly-polar substances were passing the membrane, but they
showed that on the whole, reverse osmosis was effective in re-
ducing the toxicity.
41
-------
(xl.28)
SECONDARY EFFLUENT
Dissolved organic carbon:
Initial, 9.3 ppm
Pre-flush, 6.3 ppm
CARBON - TREATED
Dissolved organic carbon:
Initial, 10 ppm
Pre-flush, 6.3 ppm
FINISHED, CHLORINATED
Dissolved organic carbon:
Initial, 6.5 ppm
Pre-flush, 5.0 ppm
Dissolved aluminum (incl.
colloidal), 0.4 ppm
SAMPLE SIZE: in each case,
2 liters
Figure 17.
VOLUME
Chromatograms of three samples of Pomona,
California wastewater, collected in May, 1978
42
-------
DENVER SEC. TREATED
MAY 1-3, 1978
(x 1.28)
E
c
cvi
*»
LJ
O
m
o:
o
(O
CD
VOLUME
REVERSE QSMDSIS REJECT
Sample volune, 850 ml
Dissolved organic carbon:
Initial, 24 ppm
Pre-flush, 8 ppn
REVERSE OSM3SIS PERCOLATE
Sanple volune, 2.0 liters
Dissolved organic carbon;
Initial, 2.2 ppn
Pre-flush, 2.0 ppn
GRADIENT
Figure 18. Denver secondary effluent, treated by reverse
osmosis in the laboratory.
43
-------
Liquid Chromatography of Tap Waters and Environmental Samples;
Solvent Blanks
Some miscellaneous chromatograms are included here. Since
the carbon content of these waters is much less than that of
wastewaters, attention to solvent blanks is necessary. Figure
19 shows the blanks with freshly redistilled water and two
commercial brands of liquid chromatography grade methanol. The
water was purified as described above. The XAD-2 resin filter
used to clean the house distilled water must be changed and re-
fluxed with methanol in the Soxhlet extractor mode every month
or so.
Figure 20 shows a chromatogram of the pristine spring water
that is described above in the section on volatile compounds.
Ionized organic compounds are present, as well as the ubiquitous
humic acids. The small peak at the end of the gradient is due
to the solvent blank. Note that full-scale deflection is only
0.08 absorbance units.
BRAND "X"
BRAND "Y"
(xQ.16)
Figure 19.
Blanks with redistilled water and liquid
chromatography-grade methanol.
44
-------
MOUNTAIN
SPRING
WATER
(1100 ML)
(xO.08)
Figure 20
GRADIENT
Chromatogram of a "pristine
mountain spring water.
45
-------
Chromatograms of Denver tap water are shown in Figure 21.
The effect of pH, previously discussed, is clearly seen, and it
is evident that this water contains considerable humic material.
The sharp, high peak beyond the end of the gradient was spurious.
It was caused by a surface-active agent of formula (CoH7) ~,C *CgH4
(0 CH2CH2)n'OH, which was coming out of the Millipore HAWP
filters we were using at that time. (Identification was made by
mass spectrometry and ultra-violet spectrometry). We changed to
polycarbonate membrane filters, and then, when we confirmed that
these filters could absorb dissolved naphthalene from water, we
avoided membrane filters altogether and used only glass-fiber
filters.
Figure 22 shows chroinatograms of Boulder tap water; Figure
23 shows water of the South Platte River in Denver, above the
sewage plant discharge, and water from a shallow well in
Boulder, 0.5 kilometers down from the old, now abandoned plant
of Arapahoe Chemical Company. These two chromatograms were
early ones obtained with the microparticulate, analytical C-18
column. They show the resolution that such columns will give,
and Figure 23b shows tho complexity of this unusual water.
Secondary Chroinatography and Peak Identification
The resolving power of the large Bondapak column is poor.
The absorbent, Bondapak Cng-Porasil B, was chosen because it
would absorb and desorb trie constituents of filtered sewage over
and over again with no irreversible sorption; moreover, it is
easy to pack and gives little back pressure. With this column
there is no hope of isolating individual compounds, save in
exceptional cases like i:he surfactant from the filters that
appeared as a spurious peak in Figure 21. Each fraction, Gl
through G5, contains many compounds of moderate molecular
weight, as well as the ever-present humic substances. To separ-
ate individual compounds and to have any hope of identifying
them, it is essential to use columns of high resolving power,
and preferably columns and solvents that have different kinds of
selectivities. The idea is to collect fractions from the first
column, concentrate them by evaporation in the manner described,
and then inject them i jr.: o I'h^ high-resolution column.
So far, three kind:-> o1: absorbents have shown promise as
packings for secondary high-resolution columns. One is a strong
base anion-exchange resin (Awinex A-25, Bio-Rad Laboratories);
using buffered alcohol-water mixtures, we have achieved good
chromatograpby of th-? P.usr> peak, which contains very polar
compounds, and have: . ,, ./olyacrylate, etc.) and different
46
-------
DENVER CITY WATER
(b)
pH8.3 I*0.64
GRADIENT
GRADIENT
Figure 21.
Chromatograms of Denver tap water
showing effect of pH.
47
-------
FLUSH
GRAD.
Figure 22.
Chromatogram of Boulder tap water, showing
effect of the addition of caffeine.
48
-------
Figure 23.
FLUSH GRAD. FLUSH GRAD.
Chromatograms of (A) of South Platte River
above sewage plant, and (B) of a highly
contaminated shallow well near Boulder.
49
-------
ANION-EXCHANGER CHROMATOGRAMS
FLUSH
PEAK
o b
END OF
GRADIENT
Figure 24.
Chromatograms of flush peaks and a weakly
polar fraction on anion-exchange resin;
eluent: acetate buffer, pH 5.5 in 25%
ethyl alcohol.
50
-------
functional groups. Materials or small, particle size, specially
designed for liquid chromatography, have become available from
Japanese manufacturers in the past year or two. The third class
is the microparticulate bonded packings that are so widely used
in liquid chromatography today. The moot popuJar of these, and
the only one we have worked with to date, 3 s the microparticu-
late C]_g-bonded silica. Chemically this is the same material
as we used in our main column, but the particle size is much
less, the available surface area is much greater, and there are
other minor differences. The loading capacity is small, but the
resolution is excellent.
Figure 25 shows chromatograms obtained on porous polymers.
The solvent in each case was 55% (v/v) acetonitrile, 0.02 molar
in tetrabutylammonium hydroxide. Curve A. shows the green G2
fraction on Hitachi 3010 polystyrene gel; curves B and C are for
G3 and G4, respectively, on Toyo Soda r,!SK LS JiO polystyrene
gel. The peaks appearing earliest are those corresponding to
the highest molecular weight. Not shown are some preliminary
data from the prepacked Toyo Soda column ol a hydrophilic gel,
G-2000-SW, which establishes the molecular weight range in
fractions Gl and G3 to be 20,000 and below. The resolution of
the green fraction, G2, into two component,'-; has been mentioned;
the first, of high molecular weight, :is a liunt brown color; the
second, which is blue-green, probably had ... molecular weight
below 1,000. Fraction G3 is clearly very complex, while frac-
tion G4 is neatly divided into a portxon of high molecular
weight that comes out fjrst, fo] lowed by a sr-rier. of compounds
of relatively low molecular weight' (b^iow 'f()f)0) that is par~
tially resolved.
We are on the threshhold oi solv.iug cue problem that has
perplexed us, the problem of removing Lne high-molecular-weight
humic substances in order to faciliLato chemical analysis of the
simpler compounds. We see three ways to appjoach this problem:
1. Absorb the humic acids or; -jr, .jnic *n-~exchange resin
column. Inspection of the res.I.ii shows that they are
retained very strongly .-n-; a marrow, dark brown band at
the entrance to the colunri.
2. Use size-exclusion porous pojyirot gels, as lust de-
scribed .
3. Filter the solutions throviqh membrane ultrafilters like
the "Diaflo" (Amicon B.V./Th^ Netherlands) ultra--
filters used to separate pi-jtoins. We have seen that
these filters can fractionate onr GJ samples.
The least polar fraction fron'. the me i a column, fraction G5,
contains very little humic substances. We therefore injected it
directly into a microparticulate1 bonded C o-silica column,
MicroBondapak-C^g (waters Associates) - I'vilh t.his column we used
the same solvent system as in the mn Lr,. coljmn, namely methanol
-------
Figure 25
Chromatograms of: (A) G2, (B) G3,
(c) G4 on porous polymers.
52
-------
and water. The procedure was the following:
The G5 fractions from 10-20 liters of secondary sewage
effluent were evaporated nearly to dryness, then methanol was
added to give 3-4 ra£ of a concentrated solution that was about
70% in methanol. Portions of this concentrate were injected
into a stream of 60% methanol by means of a valve and sample
loop; they passed into a MicroBondapak-Cig column, at the outlet
of which was an ultraviolet detector, and in some experiments a
fluorescence detector also, connected downstream from the UV
detector. The flow rate was 0.5 mJi/min. The 60% methanol was
passed for a while, generally 20 minutes, then the methanol
concentration was raised from 60% to 100% over 30 minutes
according to a linear program. The UV absorbance and fluores-
cence were recorded.
Curves like those in Figure 26 were obtained. The heights
of the peaks relative to one another varied somewhat from one
collection of sewage to another, but the positions (retention
times) were the same, and the very sharp peak near the end of
the gradient was always present.
As a first step to identification, the peaks were collected
and their ultraviolet absorption spectra and fluorescence
spectra were measured. Then the solutions from each peak of
interest were evaporated in weighed vessels at room temperature,
and the weights of the residues were found. The residues were
dissolved in appropriate solvents (carbon disulfide, carbon
tetrachloride) and the infrared spectra run. Solutions of the
residues in methanol were examined by mass spectrometry, using
heated-probe injection and also gas chromatography-mass
spectrometry.
As another aid to identification, once we had decided on a
solvent program (20 minutes isocratic flow with 60% methanol, 30
minutes linear program to 100%), we injected a number of known
compounds and noted their retention times.
The high, sharp ultraviolet peak of Figure 26 was studied
in detail. Its retention time is close to that of dibutyl
phthalate. It is obvious from visual inspection that this peak
is complex. Watching the drops of solution as they emerge from
the detector, one sees a pink color appearing just before the
sharp rise in ultraviolet absorbance. The solution collected
over the UV peak is pink, and when it is evaporated, a bright
red solid speck appears before the solvent is all gone. The
bulk of the residue is a colorless liquid. The absorption
spectrum of the solution has a peak at 547 nm, corresponding to
the red color, and a much stronger peak in the ultraviolet at
262 nm.
Mass spectrometry showed the presence of three compounds.
53
-------
I
II
FLUORESCENCE—~j[
U.V
Figure 26.
GRADIENT
Chromatogram of G5 on
MicroBondapak-C^g.
54
-------
The most prominent and the least volatile is tri(2-butoxyethyl)
phosphate, a common plasticizer. This compound does not absorb
in the ultraviolet, and it was a coincidence that it came out of
the column along with a strong ultraviolet peak. The other
compounds were phthalate esters, one being dioctyl phthalate.
Yet phthalate esters have ultraviolet absorbance maxima near
274 nm, not 262 nm. The observed 262-nanometer peak must be due
to another substance. A search of Grasselli's Atlas of Spectral
Da_ta suggests some possibilities, but we have not positively
identified this compound.
Counting the red substance, which we have also not identi-
fied, there are at least five compounds in this one sharp ultra-
violet peak. The peak that follows the high peak is also com-
plex, to judge from the mass spectrum; the mass spectrum corre-
sponds to a nonylphenol ether surfactant, yet the absorbance
maximum'is at 275 nm, which suggests a phthalate ester.
The infrared spectra were of little help. Both fractions
showed weak bands due to oxygen-carbon vibrations. The weights
of the residues were less than a milligram, and corresponded to
concentrations in the original filtered wastewater of 100 and
150 ppb (rnicrograms per liter) for the two peaks collected,
assuming no loss in. processing. One can see the magnitude of
the analytical problem, and the need for processing large
vo 1 ume s o f. wa s t ewe? t e r .
SUMMARY
We are devising methods for the chemical analysis of the
very complex mixtures of organic compounds found at trace levels
in natural waters, treated waters, and wastewaters. One part
of our problem has been essentially solved, that is the identi-
fication and measurement of volatile substances that can be
swept out of water by warming and bubbling nitrogen. These
substances include chloroform, other chlorinated hydrocarbons,
benzene arid toluene, all of them substances that are toxic to
some extent. They are trapped on a special absorbent (Tenax)
and transferred to a gas chromatograph, where they are analyzed
and identified by mass spectrometry. Concentrations of one part
per billion and less can be measured. Applying our technique to
wastewater and drinking water, we see what substances are
present and how they are affected by treatments such as chlori-
nation, ozonation, and filtration through activated carbon.
Carbon treatment: J s being introduced on a very large scale for
removing organic compounds from drinking water and wastewater;
our findings suggest that it removes toxic substances but is not
especially effective; in removing traces of volatiles.
The anaJ.ys i s of organic compounds in water that are not
volatile is a much more difficult problem. Nobody has found a
satisfactory solution, but we have made a start. We have been
55
-------
able to "strain out" or selectively absorb a part of the organic
material, using an absorbent called Bondapak-C^g, followed by
stripping the absorbed substances off the Bondapak by a series
of mixtures of water and methanol. We obtained several frac-
tions, graded them according to "polarity", or compatibility
with water, then concentrated them by evaporation so that the
substances originally present in say 10 liters of wastewater,
are recovered as a series of six fractions, each of volume 4 m£.
To guide the fractionation process we record the absorption of
ultraviolet light. The record on the chart paper lets us see
what is going on, and lets us compare the results of various
wastewater treatments.
The concentrated fractions are submitted for cellular
toxicity tests. Lately, we have also been testing samples for
mutagenic activity. Our Bondapak absorbent retains only one-
third of the total organic carbon in wastewater, but the non-
retained material does not seem to be toxic. The toxicity is
concentrated in the less polar fractions. It is here that we
would expect to find pesticides and products of industrial con-
tamination. As to mutagenic effects, we have tested four sets
of samples and the data are not consistent. In some tests the
polar fractions were found to be strongly mutagenic, particu-
larly the green G2 fraction, while the less polar fractions were
not. In another test, Gl was mutagenic but G2 had little effect,
Clearly, the composition of a municipal sewage effluent varies
from day to day, and perhaps we should not expect consistency.
Mutagenesis may be due to a minor component appearing on the
border between Gl and G2, a component whose concentration varies
greatly from day to day. Mutagenesis is a serious matter, how-
ever, and more tests will be made. The cellular toxicity of the
less polar fractions has been confirmed many times. We have
used the cellular toxicity tests to judge the effectiveness of
carbon treatment and reverse osmosis.
The next step is a more detailed chemical analysis of the
various fractions and the identification of specific chemical
compounds that are responsible for the toxicity. We are attack-
ing this problem with advanced techniques of liquid chromato-
graphy. To make progress we need to refine our separation
technique, and to collect toxic fractions from much larger
volumes of wastewater.
56
-------
SECTION 4
TOXICITY TESTS: CELLULAR METABOLIC STUDIES
INTRODUCTION
To determine the total health effects of a substance in the
environment is a formidable task. Because of the complexity of
mammalian cellular metabolism and the large variety of thousands
of interacting chemical reactions in the body, it is impractical
to test a suspected toxin against every known metabolic step or
sequence of steps. Even if this could be done, the ability of
the cell to compensate, within limits, for many types of stress
would make interpretation and extrapolation extremely difficult.
The situation is simplified if the compound being evaluated has
a high specificity for a single enzyme system. For example,
organo phosphates (Parathion) specifically inhibit acetyl-
cholinesterases and quickly disrupt the body's neuro-muscular
system. However, apart from acute toxicity of this type, which
is relatively easy to detect, we are more concerned about the
effects of chronic exposure to much less acutely toxic sub-
stances of low concentrations. Therefore, a variety of com-
pounds, including aromatics, paraffins, and their chlorinated
derivatives, were studied, as well as fractions of unknown
compositions derived from treated sewage and purified water.
We also wished to measure the relative effectiveness of water
purification steps such as charcoal treatment and reverse
osmosis with a view to developing on-line biological testing as
described in a new grant proposal.
The main difficulties of of obtaining meaninful data can be
categorized as follows:
1) Concentrations of test substances are low (from 1.0 to
less than 0.1 ppm).
2) Exposure time for rapidly metabolizing cells in vitro
is limited to 30-300 minutes to avoid the effects of
normal deterioration of cellular metabolism.
3) Of all the possible pathwavs, which are the ones to be
selected as the most valid as test models?
4) If cellular toxicity is detected in vitro, how does one
extrapolate this data to evaluate gross human toxicity?
During the period of this grant considerable progress has
been made in dealing with these difficulties, although the
57
-------
problem has not been solved in its entirety. Nevertheless, an
understanding of the short-term effects of chemicals on metabol-
ic reactions has led to a reasonably accurate prediction of
their effects in man (12) and is in general agreement with avail-
able data on gross toxicity.
CHOICE OF A METABOLIC PATHWAY
In attempting to deal with the problems of low concentra-
tion and relatively short periods of exposure, it is essential
to use cells that are extremely sensitive to a very small concen-
tration changes in their immediate environment, on the order of
10~6 - 10~5M, and which react explosively and uncontrollably
once they are triggered. Two types of cells, platelets and
leucocytes, which normally fulfill these requirements in the
body? were studied.
Normally, when a blood vessel is cut or damaged the
platelets aggregate to each other to form a plug which stops
bleeding and acts as a catalytic surface for further coagulation.
In the course of aggregation the platelet becomes degranulated,
releasing a number of substances, including ADP, serotonin and
a variety of proteins. Platelet function is linked closely to
adenylate metabolism (13) and is regulated by membrane function.
Consequently, lipid soluble chemicals which change membrane
function can affect platelet metabolism and behavior. This
concept is important not only from the viewpoint of hemostasis,
but is also related to the role of platelets in forming unwanted
plugs in the lumen of arterial vessels, thus contributing to
thrombotic coronary disease, a major cause of death.
Another type of cell, which also functions explosively to
stimulus, is the white cell in the blood which forms one of the
defenses against infection and malignancy. These cells can
ingest or phagocytose bacteria and kill them with hydrogen
peroxide generated metabolic action. Thus, substances which
impair these functions or cause these cells to inappropriately
attack the tissues of the host are of concern. The white cells,
like the platelets, generate and use large quantities of high-
energy-transducing purines (ATP, etc) and have efficient path-
ways for salvaging ATP from degradation products such as
hypoxanthine (HYPX).
The third difficulty of choosing a relevant pathway to
study is not as easily solved and all choices of metabolic
pathways are open to some sort of criticism. However, there is
one sequence which is of central importance to both the ener-
gizing and kinetic control of all metabolic steps. This system
is the adenylate control system which has come to be recognized
as being of prime importance in energy transduction, storage and
control of metabolic rates and directionality (14). The import-
ance of the fast-acting adenylate control system cannot be
58
-------
exaggerated for a large number of chemical reaction sequences
which typify living organisms. The adenide nucleotides interact
with all sequences in a complex living cell and uncontrolled
changes in the relative concentrations of ATP, ADP and AMP would
adversely affect the rates of all metabolic reactions and thus
be highly disruptive. Therefore, we have developed methods to
assay adenylate pools when platelets and white cells are brought
into contact with a variety of chemicals.
The basis of adenylate control resides in the fact that
these purine nucleotides combine with enzymes and alter the
rate and direction of important control enzymes. For example,
under stress the AMP/ADP ratio is increased and stimulates the
activity of phosphofructokinase (PFK), thus accelerating
glycolysis and stimulating ATP production. Isocitrate dehy-
drogenase is similarly stimulated by AMP, resulting in an
accelerated Krebs cycle, with the ultimate regeneration of more
ATP. These reactions are appropriate for survival since a
significant fall in the concentration of ATP leads to increased
AMP by the pathways ATP, ADP, AMP; the AMP mediates correction
of the ATP level by stimulating glycolysis and phosphorylation.
Similarly, when the demands on the cell are reduced, PFK is
inhibited and reserves are conserved. Reactions of this type
are normally taking place every instant with the result that
ATP levels are maintained at optimal levels.
Another important aspect of this system is that evolutionary
design has favored a high equilibrium constant over a large
yield of ATP. A high.equilibrium constant allows the cell to
advantageously use very small amounts of food while paying the
price of a reduced energy yield. A higher yield would require
a plentiful supply of fuel at all times, and this would reduce
the potential for survival. However, since the efficiency of
ATP production is relatively low, stressors and toxins have the
effect of further reducing the ATP available for useful work and
thus seriously reducing the cells' ability to carry out their
functions. This is most readily seen in platelets which are
unable to make ATP "de novo' from its building blocks but must
rely on scavenging and recycling the purine ring compounds such
as adenine and hypoxanthine from its surrounding.
It is remarkable that the role of the adenine nucleotide
pool is unique in being involved in virtually every metabolic
sequence in the cell. The role of the adenylates is not specific
to any single pathway but, more than any other compounds, they
couple and correlate all the metabolic activities of the cell,
giving rise to biological homeostasis and function. For these
reasons, the purine nucleotide system was chosen for our study
of health effects of substances in drinking water. An outline
of the metabolic pathways utilized in this project is shown in
Figure 27. The effect of stress on this pathway is to reduce
ATP and increase AMP and hypoxanthine (Figure 28).
59
-------
ATP =s ADP :=±=AMP
HYPX INOS IMP
PARTIAL PURINE NUCLEOTIDE
PATHWAY
Figure 27. Partial purine pathway.
SALVAGE
CELL MEMBRANE
HYPOXANTHINE -
XANTH.
Figure 28. Effect of stress on ATP metabolism.
60
-------
METHODOLOGY
Because of the nature of the experimental conditions in
which radioactive adenine was used as a precursor, the only
radioactive compounds formed in significant quantities were ATP,
ADP and AMP, with occasional formation of inosine monophosphate
(IMP and HYPX). Dose response curves were used to express the
results of known toxic compounds on adenylate metabolism as well
as the effects of sewage, tap water and water effluents during
treatment. In outline, the platelets were first obtained as a
suspension by the standard method of differential centrifugation
of anti-coagulated blood. Care was taken to prevent stressing
the platelets during blood drawing or subsequently in the labor-
atory. Thus, only a 'clean' venipuncture from healthy blood
donors was used and high speed centrifugation and washing avoid-
ed. Suspensions of white cells were prepared and treated
similarly.
In a typical experiment, 5p£ of U-14-C adenine (60,000 cpm)
was added to a 0.5 m£ suspension of platelets (2-5xlO~Ycu mm)
or white cells (10,000/cu mm). Ten microliters of the test
fractions to be evaluated were added. After incubation at 37°C
for 30 minutes the cells were cooled in ice and centrifuged at
3000 rpm for 3 minutes. The supernatant was removed and 0.2 m£
of cold 13% perchloric acid (PCA) was added to the pellet.
After mixing and centrifuging, the PCA supernatant was analyzed
by thin layer chromatography (TLC) and high performance liquid
chromatographic systems (HPLC) described below. The TLC system
allows one to measure the kinetic synthetic ability of the cells
to convert U-14-C adenine to U-14-C purine nucleotides. The
HPLC system measures the total pools of these substances.
Thin Layer Chromatography (TLC)
Ten microliters of the PCA supernatant were applied without
heat to a 2-centimeter strip of an Eastman Kodak Cellulose Thin
Layer Chromatogram #13255 with a mylar backing, and developed
with a mixture of water: formic acid: tertamyl alcohol:
1:2:3 for 5 hours. The chromatogram was dried in air, cut into
15, 1-centimeter strips, and each placed in a 5-milliliter count-
ing vial with 3 m£ of scintillation fluid. The vials were count-
ed in a scintillation spectrometer and each count calculated as
a percentage of the total U-14-C nucleotide pool. A typical
representation of the toxic effects of m-xylene is seen in
Figure 29 with substantial decreases in ATP and ADP and an in-
crease in AMP. Significantly greater amounts of unmetabolized
adenine are seen in the cells exposed to m-xylene. Recovery is
95-102%. In a large series of normal platelets carried out in
connection with other projects ATP was 65-70%, ADP was 10-14%
and AMP was 1-5%. Changes of 15% or larger in ATP levels were
regarded as being above the analytical and biological noise
levels for this determination.
61
-------
30-
O
u.
O
10 H
ATP
TLC
Cells Alone
Plus K) ppm *
m-xylene^ i
I
Adenine
10
15
CM
Figure 29.
Toxic effects of m-xylene
on platelets.
High Performance Liquid Chromatography (HPLC)
The major technical advance which makes it feasible to
consider on-line capability is the use of high performance
liquid chromatography (HPLC) for the rapid determination of the
pool of adenine nucleotides at the rate of 15 minutes per
sample. The peak heights are measured and used for calculation
of pool size. Both TLC and HPLC are sensitive down to the 10 M
range (15) .
A Waters & Associates HPLC apparatus with a U-18 micro-
Bondapak reversed phase column was used. Fifteen microlilters
extract was injected and 0.1 M KH^PC^ was pumped through the
column at 1.5 m£/minute. The nucleotides and related compounds
were adequately separated in 12-15 minutes and were detected by
UV absorption at 254 nm (see Figure 30). Identification was
62
-------
ATP
t
E
c
in
CJ
co
z
UJ
o
o
H
Q.
O
PCA
START
ADP
AMP
ADENINE
6 9
TIME(MIN)
12
15
Figure 30. Elution curve of pure standards
using HPLC.
63
-------
obtained using retention time and the addition of known com-
pounds to the test PCA supernatant. In the future it is planned
to pump the effluent from the HPLC system into a flow-through
scintillation spectrometer so that specific activity of each
compound and be computed, thus obviating the need for TLC.
Recovery and reproducibility of the standards was 95-103%. The
peak height was found to be linearly proportioned to concentra-
tion, regardless of peak width, in the range 10~4 to 10~9 moles
or purine compound per sample injected.
RESULTS
The results are divided into three main sections:
1) Influence of known toxic substances on platelets,
neutrophils, and monocytes.
2) Effects of concentrated fractions of water on neutro-
phils and monocytes.
3) Study of reverse osmosis permeate and carbon treatment.
The chemical fractionation of various waters was carried
out by Dr. Walton and is fully described in his section of this
report. In order to avoid drawing up numerous tables, graphical
representation was used whenever possible and the numbers
printed on the graphs.
A list of the substances tested is shown in Table 7. For
comparison of the cellular results with gross toxicity, the
Repository of Toxic Effects of Chemical Substances (RTECS,
compiled by NIOSH) was used to compile the exposure standards
and toxicities of these compounds (Table 8).
Effects of Known Substances
Known compounds with the highest available grade of purity
were added to the cell suspensions as described above. Initial-
ly, measurements were made of U-14-C adenine incorporation into
ATP, ADP, and AMP, using TLC. Later on, when HPLC became avail-
able, the total adenylate pool was measured. A control with no
additive was run with each batch of cells. Metabolic stress
effects are indicated, either by a reduction in the 14-C or 12-C
ATP pools (Table 9). Platelets, neutrophils, and monocytes were
used for testing the toxicity of the known substances and are
discussed below.
Platelets--
Figure 31 shows significant reductions from normal for
U-14-C ATP in platelets treated with a range of chemicals, all
at 2 ppm. In extrapolating from these results to gross toxicity
data available in RTECS, we found that certain correlations
exist. For instance, although the correlation is not perfect,
chloroform is more toxic than toluene in both the cellular ATP
64
-------
TABLE 7. KNOWN SUBSTANCES TESTED ON THE 3 CELL TYPES
Cell Type
Test Substance Platelets Neutrophils Monocytes
Chloroform + + +
Toluene + + +
1:1:1 Trichloroethane +
1:2 Dichloroethane +
o-dichlorobenzene +
Tetrachloroethylene + +
Hexanal + +
m-Xylene + +
p-Xylene +
Methanol +
n-Octane + +
Trichlorethylene + +
N-Hexane +
65
-------
TABLE 8. GROSS TOXICITY DATA FOR HUMAN AND RAT EXPOSURE1
Substance
Human
Work Std. (ppm)
(NIOSH)
LD50
mg/kg (rat)
Chloroform
0-dichlorobenzene
p-dichlorobenzene
DimethyIsulfide
1,2-Dichloroethane
Carbon tetrachloride
Trichlore thylene
Tetrachloroethylene
Hexanal
Toluene
Ethylene dichloride
o,m-, & p-xylene
Trichloroethane
n-octane
n-Hexane
10
50
75
50
100
100
200
200
350
500
500
800
520
500
535
680
1770
4920
5671
4890
5000
5750
(dog)
5000
9470
Repository of Toxic Effects of Chemical Substances, NIOSH USDPH
Rockville, Md., 1976.
66
-------
1
u
p§
En
O
EH
Q
£3
H
O IS
EH H
CO Q
Ea &
r-3 O
pq ftj
EH En
KH1 Q
ft W
Q H
IS -*•!
rO P i
CD
^3
13
<
•H -H -H -H
4^ x3 xj -P x:
Cij P-* P.I d) CX
O O O H O
^.-i Lj Lj QJ C i
-P -P -P -p -P
3 |3 3 (0 3
(DO) QJ rH 0)
^i IS £S ft S
CO CN i — 1 LO ^J* CTi CN ^D *sF VO ^O LO C5 LO LO f*^ P^-
rHM'CNnrH OCNCNO OOC^OOO '3'M'CN O
rH rH rH rH
OrHOO OOOOO OOOfO OOOCN OOOOOCTi
O^OO CNOOOO rHrOV£)LT) r-rOrHVD CTiOLnOOr-~CN
in VD ^DU3i^5 IDVDi^J LDV£>'*CNrH
rH O O 0
OrHOO OOrHOO OrHOO OrHOO OrHOO
OOrHO OOOrHO OOrHO OOrHO OOrHrHCMO
rH rH rH r-H rH
Q)
. 1
r~i
f~|
-P
Q) 0) 6 — -
O C SH -a
!H 0) o Q)
rH O 0) H MH 3
O rH C >1 O fi
£ ,£ Q) X M -H
(0 U 3 ffl O -P
X -H rH -P rH £
OJ 5-1 O 0) ,d O
S3 EH EH g U O
67
-------
^^
T3
0)
3
a
•H
-P
c
o
D
CA
W
^
PQ
<
EH
rH
H
Q)
U
ft
r i
c*
< g
o
T
r i4— i i i
*^^l ~ *T-*
hi O X!
ft
ffi + X
(d
CM (D
rH QJ
t_>
<44
O
i
ft O
Ur~ , /•>
bH U
hi < QJ
EH
•531 «a<
rH rH
U U
o\o
•
o
G
o
0 g
Pj
>-iJ
13 QJ
0)
T3
T3
«rf*
P-N
-P 4-1 4-1 rH H
Q) QJ 0) -H -H
H rH rH X! X!
(U (U 0) QJ Q
4-1 4-1 4-1 O 0
(d (d fd ^ !_
H rH rH 4-> 4J
ft ft ft 33
01 QJ
S IS
o^ooco oino^o o^r^in
fOrHrHO rO^rr^lX) OOrHCNO
CNCNCNCN CMfNrHrH CNCNCNCN
OO OOO OOO 00 C
lir-cri in^r-H i^jooo'sf CNPO c\
r>- r- vor^vo r-^r~ r~-vo t~-
OrHOO OrHOO OrHOO OO C
OOrHO OOHO OOrHO OO C
rH rH rH CN
QJ
*s- rr~l r*
r^^~J K
X -H QJ
O O rH
XJ , rd
4-i X X!
0) O O 4->
g -H x; x!
1C 4J QJ
CN O QJ (d
•H g C
QJ ^1 Q) 1
G QJ ^> C >i
QJ O QJ X QJ Q
rH ^ OH O G C
rd QJ g (d x! rd n
XJ O O XJ 4-1 4-1 >
4J W SH 4-J Q) O Q
Qj-H X! QJ g OX
rd I I rd I II
& vo CN Gl CM G C
rH rH
•H -H
XI X!
J QJ QJ
o o
^ M
4-> 4-1
3 3
QJ QJ
13 13
i ill i i i i
i ill i i i i
>O OOO OOOO
in or-~-ro rHco^cN
CN t--r— IH r~~kDU5r~~
OOO O rH O O
)O OrHO O O i— i O
>O OOH OOOiH
CN
QJ
G
(d
X!
4-> QJ
Q) G
O QJ
^ N
0 C —
H QJ T3
X! X) QJ
J 0 0 3
5 -H M G
3 M O -H
' 4-> rH -p
j i x; c
5 H U 0
H -H U
5 H Q —
68
-------
S-*
TJ
(!)
3
G
•H
-P
C
O
o
• — •
en
W
r}
CQ
<
EH
rH
rH
CU
CJ
CM
EH
< e
U
T
r )L_ i i i
\^t ' i -r-*
r3 U.C
CM
ffi + ^
rrt
iu
CN CU
, i /-»
1 1 i-iH
U
m
0
rH
CM O
Uf-t <~>
c^ w
*i <$ a.
EH
rtf T
rH rH
CJ 0
0\°
•
u
c
o
u e
o.
>-M
T3 CX
0)
13
3
<3
-P rH
0) -H
i r-i
^n ri-i
cu ex
-P O
to JH
rH ^J
CM 3
CU
S
CN ro r~ oo
r^ in rH o o
rH rH rH rH
O O O OO
^D CN 1 H in
r^ *£> in
000
O -^f T O rH
O rH rH O rH
>,
to
&> tO
fi 'O
•HO) C
-X -^ „ O
fi tO , U fO
•H to cd CU C
M -P uj c/1 O
T3 0 3= g
cu ... TO
•H o 2 ^ ^
4J G
tfl 0 -o C -P
fi U QJ 0 G
G -H CU
•H H 'a. -1-1 ^
O p-; Mm
U ^ "^ UJ h W
•p -p
CU (U
M-,1
r— i
0) CU
-p -p
(0 (0
rH rH
CM CM
o vo oo o VD m
H O O ro oo ro
H rH rH rH rH rH
O O O O O O
in ro m vo T vo
in in sr in m in
•if? in o o o o
o CN in o o o
rH CN O O
VO CN
rH
CU
H3
•H
tn MH
rH rH
O 3
M to
-p
G rH CU
O >i CO
U A O
4-> 0
CU CU 3
> g H
•H -H O
-P Q
fO x-.
& ^-. -H
CU -H -H
& ^ ^
69
-------
response and the qualifying toxic dosages. Similarly, octane is
less toxic than toluene.
O
O
Q.
100
80
U 60
u.
O
O
o
U)
o:
jj*
20
0
88
62
65
46
32
33
12
OCTANE
TOLUENE
TETRA
CHLORO
ETHYLENE
p-XYLENE TRICHLORO fl-HEXANE
ETHYLENE
INCREASING TOXICITY
CHLOROFORM
Figure 31.
Effects of known substances at
2 ppm on U-14-C ATP pools of
platelets using TLC.
Dimethyl sulfoxide (DMSO) and glucose were included as neg-
ative controls. DMSO is used for cell and organ preservation and
has been given in large amounts to patients with scleroderma.
DMSO is also used to solubilize the constituents of water con-
centrates. Glucose is also ingested in large quantities in our
diet and is not acutely toxic to cells over a large range in
concentration (600-1200 ppm) unless insulin or glucogen metabo-
lism is disturbed. Our results showed that these two materials
were without significant effect on the metabolic parameters being
70
-------
used as criteria for toxicity (Table 9), thus lending increased
validity to the usefulness of these tests for establishing
toxicity criteria.
Although there are many factors which influence the toxic
effects of a chemical, the broad correlation which is present
indicates that cellular ATP measurements may not be too far
removed from clinical toxicity effects. Although this work was
concerned with ATP because of the universality of its function
in living systems, one need not be restricted to this pathway
in future work.
Neutrophils--
Rankings of the toxic effects of the chemicals acting on
neutrophils at the dosage levels of 0, 0.1, 1.0, and 10 ppm of
substance'added is shown in Figures 32, 33, 34, and 35. As with
platelets, neutrophils respond to varying concentrations of
these compounds, still preserving the main features of the RTECS
ranking. The dose response characteristics are shown in Figure
35. Chloroform and trichloroethane give significant responses
at 0.1 ppm, whereas toluene, xylene, 1:2 dichloroethane and
o-dichlorobenzene affect these cells at the 1 ppm level.
I00r
O
O
a,
a,
-------
o
(0
to
UJ
z
UJ
Ex
UJ
z
UJ
2
a:
2
o
I2
a _
770$
77 OCH
— HUJ
Ouj
o
QCO
oz
CM _J<
OUJ
UJ
a:
o
O
O
O
00
O
10
O
CVJ
o
CM
i
OJ X!
O
C en
fO H
-P -H
P O
to M
-P
? CD
O C
c
o
O <—1
o
-P O
U P.:
0)
m a,
M-i EH
w <;
ro
CD
HOOd dlV
dO
%
72
-------
100
o
0
0- 80
CL
1-
2 60
O
u.
o
o 40
u
o
Ul
tr 20
n
_^
—
—
19
95-98
89
METHANOL m-XYLENE
CHLOROFORM
TOLUENE
HEXANAL
TETRACHLORO
ETHYLENE
INCREASING TOXICITY
Figure 34.
Effect of known substances at 10 ppm
on the U-14-C ATP pool of neutrophils
by TLC.
73
-------
CHLOROFORM
TOLUENE
m-XYLENE
O.I 1.0 lOppm
\-2 DICHLORO-
ETHANE
O.I 1.0 lOppm
lOppm
0-DICHLORO-
BENZENE
lOppm
TRICHLORO -
ETHANE
O.I 1.0 lOppm
O.I I.O lOppm
Figure 35. Dose response curves of pure
substances acting on neutrophils,
74
-------
Monocytes—
Monocytes were allowed to attack red cells labeled with
radioactive chromium and the killing power was determined by
measuring the chromium release. The same cells were examined
for changes in U-14-C ATP and total ATP pool using both TLC and
HPLC. The results were compared with each other at 0.1, 1.0,
and 10 ppm of added known substances (Figures 36, 37, 38, 39).
In general, the TLC and HPLC dose response curves were in good
agreement (Figures 39,40). The HPLC curves often showed a more
sensitive response at the 0.1 and 1 ppm levels. The monocyte,
however, was found to be most resistant to chloroform at concen-
trations of 0.1 and 1.0 ppm, responding at 10 ppm. Except for
tetrachloroethylene the monocyte did not give a graded response
for target cell release of chromium with increasing concentra-
tion. Thin layer chromatography and HPLC were more generally
sensitive in detecting graded cell responses. These results
were most satisfactory from a technical viewpoint and large
numbers could be processed per day. However, interpretation
is more difficult with regard to RTECS ranking because of the
refractory behavior of the monocytes to 1 ppm of chloroform
which ranks high in toxicity and because of their larger re-
sponse to compounds such as m-xylenes and p-xylene~, which have
lower overall toxicities than chloroform. Nevertheless, it is
important to realize that a chemical which is relatively low in
gross human toxicity may have a considerable effect on the
activity of this important cell. In all the cells a concentra-
tion of added substance at 10 ppm was too high for the discrimi-
nation between individual compounds, which is necessary for
ranking. Ranking was more efficient at the 1 ppm level of
exposure.
Effects of Concentrated Water Fractions on Neutrophils and
Monocytes
Concentrated fractions were produced and labeled from Gl to
G9. The higher the G number the lower the polarity of the
constituents of the fraction. Figure 34 shows the response of
neutrophils to dilutions of each fraction. Due to the un-
availability of reliable total organic values, an absolute
concentration is not given for each fraction. Nevertheless,
very significant responses were obtained with toxicity increas-
ing with the G number of the fraction. Thus, G2 is seen to be
relatively non-toxic to neutrophils, but G4, G5 and G9 produce
significant reductions in the U-14-C ATP pool.
With regard to monocytes very similar results were obtained
compared to those obtained on neutrophils (Figure 41). These
results on extracts of Pomona, California wastewater note the
changes in chromium release and U-14-C ATP pools. As before,
toxicity generally increases with G number. Of special interest
is G4 and G4A; G4A is charcoal-treated in the plant and by
comparison with the untreated G4 fraction it is seen that
75
-------
UJ
look
UJ
1
— J
UJ
^v*
tt.
* 80
S
2
2 60
I
O
0
Z 40
<
a.
h-
< 20
z
z
0 0
1-
z>
- HPLC
—
1
33 34
TLC n 23 n r
C" 14 '4
43 r^JI r 3 1 5, S
P^I n r n IP_ ^ TO
J ^ 1 TOLUENE !
C 1— •
b HEXANAL TT m-XYLENE p-XYLE
CHLOROFORM TETRA
^j CHLOROETHYLENE OCTANE
Figure 36. Effects of known substances at 0.1 ppm
on the U-14-C ATP and total ATP pools
in monocytes.
76
-------
100-
80 -
o. u
H V)
z j
LJ
Z QC
2 2 40
h~ -^
0 0
IT K 20
^ 0
68
6!
41
Cr
TLC
HPLC
r
12
m
4J 1
r
U
36
^
J
30
34
22
4
n
39
9
1
\i
21 24
12
i r
—
1 2 m-XYLENE OCTANE TETRA(
6 4
CHLOROFORM HEXANAL
TOLUENE
ETHYLENE
p-XYLENE
Figure 37. Effects of known substances at 1.0 ppm on
the chromium release, U-14-C and total
ATP pools in monocytes.
77
-------
UJ
UJ
-J
UJ
tr
5
o
(T
I
O
O
z
u_
O
z
o
\~
o
D
o
u
QC
100-
80-
60-
40-
20-
0
93-100
84-100
89-91
HPLC-
TLC
Cr
CHLOROFORM;TOLUENE;
HEXANAL; p-XYLENE
m-XYLENE; OCTANE;
TETRACHLOROETHYLENE
Figure 38.
Effects of known substances
at 10 ppm on the chromium
release, U-14-C ATP and total
ATP pools in monocytes.
78
-------
CHLOROFORM TOLUENE OCTANE
HEXANAL
O.I 1.0 lOppm MI 1.0 lOppm 1 I.O lOppm 4 I.O lOppm
lOOi- ^lOOr- rOOr-
Q.
h-
01 I.O lOppm O.I I.O lOppm O.I I.O IOppmO.1 I.O lOppm
100,- ^lOOr- ^lOOr- HOOr-
6 6
IOppmO.1 I.O lOppm O.I I.O IOppmO.1 I.O lOppm
Figure 39. Dose response curves of pure substances
in contact with a suspension of mono-
cytes, 0-10 ppm. (continued)
79
-------
m-XYLENE p-XYLENE TETRA
LU
<
UJ
_l
LJ
v.
O
lOOr-
CHLORO
ETHYL-
ENE
0
-------
J I L
95
73
74
I
*_ A x
100 10
.* *.
100 10
100
100
Figure 40.
Dose response of the U-14-C ATP pool to con-
centrated extracts of sewage; TLC polarity
of the constituents decreases as G increases.
81
-------
o
;UJ
CD<
OI-O
X,
(/)
13
o
ID
0
IO
0
0
to
C\J
o
o
00
00
0
o
00
GD
00
CO
O
C\J
M
coL
tor
roL
u
i
IT)
K,
O
00
0
IO
O
CM
ao
JO
•H •
C w
^ OJ
o -u
m >i
•H 0
rH O
(13 C
C G
O -H
g to
o ^
to
O g
•H 0
-P -H
u e
rtS O
M M
OJ •
to tu
(d co
(U 3
iH
H -P
•H
-------
toxicity is greatly reduced by this treatment. Thus, it may
become feasible to monitor separate unit processes within a
treatment plant by means of our tests and pinpoint the entrance
and removal of toxic materials.
Reverse Osmosis
Our toxicity methodology is readily adaptable to the study
of pilot plant or laboratory scale purification processes. In
this regard G3 and G4 concentrates were obtained from laboratory
reverse osmosis apparatus. Both the permeate and reject of
sewage were tested using neutrophils. As before, using the
criteria relating to ATP metabolism, G3 permeate was less toxic
than G4 and the reject, as expected, was highly toxic (Figure
42) .
0.
H
0
2
PLC
0
^•K
77
m^f^**m
97
=^i
,
REJECT
— »J
98
G3
G4
G3+ G4
Figure 42.
Reverse osmosis; toxicity of G3
and G4 permeate and reject
fractions using neutrophils.
83
-------
DISCUSSION
In the space of three years we have progressed from a
theoretical understanding of stress effects in human cells to a
practical application to water extracts and the effects of unit
processes such as charcoal treatment and reverse osmosis.
The utilization of cell incubation followed by HPLC and TLC
analysis offers considerable capability for the sensitive and
rapid detection of cellular toxicity. The system is flexible in
that it is not limited to any particular metabolic pathway or
cell type and is relatively inexpensive. The system is vulner-
able in that it needs a constant supply of normal cells of organ-
isms. However, continuous 5-point dose response curves can be
obtained using platelets from 10 m£ of blood per day available
from a nearby blood bank. Alternatively, one could use in vitro
cell cultures of fibroblasts or bacteria which can be made up in
large batches. As long as sufficient cells are present an
accurate cell count is not needed, provided the same suspension
is used for test and control. If necessary, a portion of the
cells can be diverted for functional studies or other investiga-
tions .
Progress was also made with regard to three previously
mentioned problems of sensitivity of the assay, limited exposure
time, and the selection of a model pathway. In general, sensi-
tivity was adequate down to the 0.1 ppm range for known com-
pounds, and water concentrates of approximately 800 fold. It is
estimated that by increasing the ratio of test substance to
cells, an improvement of at least 10-fold could certainly be
obtained and longer times of incubation could probably improve
this by ah addition factor of 2-5. With regard to the selection
of a metabolic pathway, it is true that certain types of toxic-
ity would be missed., for example, mutagenicity and very specific
enzyme effects. Nevertheless, the adenylate pathway is of cen-
tral and unique importance in the transduction of energy and the
kinetic control of metabolism. Thus, a major disturbance here
would have serious consequences for any cell. More specific
tests of toxicity could be done in response to a detailed
knowledge of the chemical constitution of the components of the
water concentrate.
In general, reductions in ATP as determined by HPLC and TLC
are detectable for most compounds present at concentrations of
0.1 ppm, with larger effects found at 1.0 ppm. At this order of
concentration it was possible to rank the substances in order of
cellular toxicity. At the higher exposures of 10 ppm the cellu-
lar toxicity was so great as to render distinction between com-
pounds virtually impossible.
Platelets have their limitations with respect to the fact
that they lack a nucleus and its DNA. For this reason, other
84
-------
cells need to be used when mutagenic characteristics are of
interest. However, no serious problems are forseen for the
practical adaptation of these cells for continuous surveillance
of metabolic toxicity of water effluents. The availability of
platelets from small amounts of blood (1-5 m£/dose response
curve), their stability and their quick response to perturba-
tion makes them extremely suitable for our purpose. They are
remarkably stable if kept suspended in their original plasma and
not subjected to mechanical trauma. White cells (neutrophils)
were more prone to deterioration due to technical reasons, and
preparations had to be discarded on several occasions. Mono-
cytes were very stable but their responsiveness to toxic materi-
als present in concentrations below 10 ppm was not great.
Furthermore, the toxicity ranking differed from the gross human
and rat toxicity found in RTECS listings. Nevertheless, the
monocyte needs to be developed further because if its sensi-
tivity can be improved, it has the potential for detecting
toxicity in compounds that are presently not thought to be
highly dangerous.
It is difficult to decide which of these three cell models
is the most useful for monitoring purposes. They all have their
advantages and disadvantages. With on-line operation ultimately
of prime importance, the platelet is recommended because of its'
stability and ease of analysis with only small amounts of blood
(1-5 m£) needed per dose response curve. On the other hand,
neutrophils and monocytes tell us more about the immune defense
system, while platelets do not. However, larger volumes of
blood on the order of 50 m£ are required and the behavior of the
neutrophils may be subject to seasonal variations. The mono-
cytes were less responsive overall and their toxicity ranking
did not seem to follow the RTECS listings as well as the other
cells. However, monocytes gave highly reproducible results.
In the first year considerable work was expended in sim-
plifying the methodology, testing known compounds, and learning
how to control the sources of variation. In the second year
known substances continued to be tested, but the greatest empha-
sis was placed on testing fractions of treated sewage. In the
third year a high performance liquid chromatography apparatus
was purchased and it was then possible to include total adeny-
late pools in our toxicity assays as suggested by previous EPA
reviews of our work. Many more fractions were analyzed and the
results extended to a study of treated wastewaters from Pomona,
California waste treatment plant. Distinct differences were
noted in the effects of this water before and after charcoal
treatment. Both toxic and non-toxic fractions were found in
the samples. Thus, the methodology had sufficient range to
distinguish biologically between fractions of different chemical
polarity. The humic acid fraction was non-toxic, but less
polar compounds had adverse effects on the cells.
85
-------
SUMMARY AND CONCLUSIONS
The main objectives of the testing were met with regard to
the items listed below:
1) Several model systems for measuring cellular toxicity
were developed using platelets, neutrophils, and mono-
cytes.
2) Testing included known substances, water fractions, and
concentrates from sewage and water treatment plants.
Degrees of toxicity and the effects of treatment were
measured.
3} The metabolic, pathway of adenylate metabolism was found
to be suitable as a rapid method of analysis and allows
for a unified, theoretical approach to toxicity of a
wide variety of compounds.
4) Correlation of cellular toxicity with gross toxicity
quoted in the literature was satisfactory for platelets
and neutrophils.
86
-------
SECTION 5
TOXICITY TESTS: CELLULAR BACTERICIDAL STUDIES
INTRODUCTION
The objectives of the work reported in this section can be
summarized as follows: to develop functional assays that will
indicate whether the specialized activities of white blood cells
have been compromised by exposure to environmental substances.
White cells were selected for study because of their impor-
tant function in protecting the body against invasion by bacteria
and in combating malignancy. The cells carry out their bacte-
riocidal functions by first sensing the presence o.f the bacteria
moving to the site of infection, ingesting the bacteria (phago-
cytosis) , and then killing them chemically by metabolic action.
Emphasis was placed on the phagocytosis and killing power assays
of neutrophils and monocytes. Difficulties related to contami-
nating red cell removal were overcome by a modified lysis tech-
nique of short exposure to ammonium chloride; further stabili-
zation was effected by suspending the cells in tissue culture
fluid RPMI-1640 to which 10% fetal calf serum was added.
Despite these modifications the screening of large numbers
of samples was impossible since the complexity of the neutrophil
technique allowed for only 3 or 4 assays per day. This problem
became particularly acute as Drs. Walton and Eiceman improved
their techniques of separation of compounds from water, result-
ing in a large number of peaks to be tested from an individual
water sample. In order to do dose response curves on individual
fractions, this technique severely limited the number of samples
that could be analyzed at one time. Fractions from a single
water sample could be analyzed at one time. Fractions from a
single water sample would have to be analyzed over a two week
period, and the sensitivity of the bacterial assays for phago-
cytosis and killing power were relatively low. Fortunately this
problem was overcome by the use of human peripheral blood mono-
cytes in a new assay of human cellular function.
The monocyte assay was first introduced in late 1976 (16)
and was given the name antibody dependent cellular cytotoxicity
(ADCC). It required both human antibody and human cells to
destroy chromium-labeled target cells. It was soon discovered
that ADCC was an essential function in the elimination of cancer
87
-------
cells and the termination of viral infections. The human ADCC
system has the major advantage of being able to screen several
hundred samples per day. An outline of the methodology used for
the neutrophil and monocyte assays is given below. Table 10
shows the substances tested.
TABLE 10. KNOWN SUBSTANCES TESTED BY NEUTROPHIL
AND MONOCYTE ACTIVITY
1:1:1 Trichloroethane
1:2 Dichloroethane
Tetrachloroethylene
o-Dichlorobenzene
Chloroform
Dimethyl sulfide
Trichloroethylene
n-Octane
m-Xylene
p-Xylene
n-Hexanal
Toluene
Carbon tetrachloride
METHODOLOGY AND RESULTS
Known Substances
Neutrophil Testing—
The neutrophils were obtained by separating anti-coagulated
blood immediately after venipuncture using a column of methyl
cellulose hypaque. Contaminating red cells were lysed with
ammonium chloride (0.85 gm/100m£).
88
-------
After obtaining the white blood cell suspension and adjust-
ing the cells to the desired concentration of approximately
10,000/mm , the cells were incubated for 30 minutes at 37° in
the presence of the test compounds. The cells were then washed
twice by centrifugation and resuspended in tissue culture medium
RPMI-1640 with antibiotics and 10% fetal calf serum, and assayed
for phagocytosis and killing power. A portion of the cells was
immediately taken to Dr. Solomons' laboratory in the same build-
ing for ATP metabolism studies.
In order to measure phagocytosis and killing power the
cells were exposed to a standarized number of bacteria (staph-
ylococcus aureus 502A) and opsonis (8% AB serum) and incubated
for another two hours at 37° on the tilting table. This interval
of time allowed for phagocytosis and killing of the bacteria by
the neutrophils (PMN). After the incubation period the total
number of bacteria surviving the experiment was determined by
plating, and expressed as a percent of the number of bacteria
present at the beginning of the two hour phagocytic assay. This
value represents the interaction of the phagocytic process by
PMN's and the killing of the bacteria by intracellular bacteri-
cidal mechanisms. Intracellular surviving bacteria were also
determined in each assay using lysoraphin to eliminate cellular
bacteria. This value correlates inversely with the PMN bacteri-
cidal activity. Finally, the number of extracellular bacteria
(EB) was calculated for each experiment. The EB value corre-
lates inversely with the phagocytic uptake of bacteria by the
PMN's. Thus, phagocytosis is determined as the percentage of
bacteria remaining outside the cells after exposure to the neu-
trophils. Killing power is calculated as the percentage of
intracellular bacteria surviving after ingestion by human neu-
trophils .
Instead of using bacteria, an assay using the phagocytosis
of oil particles was developed in order to improve the sensitiv-
ity quantitation and speed of analysis. Oil red-o phagocytosis
involves the internalization of di-iso-decyl phthalate in oil
which is coated with E. Coli lipopolysaccharide and serum to
optimize its ingestion. The oil particle contains the dye Red-o,
which can be quantitated by spectrophotometry. Although this
method is sensitive and correlated well with bacterial phago-
cytosis, it requires too much blood to be practical for daily
use.
Dichlorethane, tetrachloroethylene, and trichloroethane at
10 ppm significantly reduced neutrophil phagocytosis (Figure 43).
Insignificant effects were observed with water fraction concen-
trates, Gl, G2, G3, and G4, but G5, G8, and G9 showed measurable
decreases in phagocytosis (Figure 43). However, the rest of the
known substances had little effect. It was also thought that
the additive could kill bacteria independently of the neutrophil,
thus making interpretation difficult. Reverse osmosis permeate
89
-------
and reject did not have any significant effect on phagocytosis
and killing power although marked effects were observed on ATP
metabolism. Because of the relative insensitivity of this assay
the monocyte assay system was developed to replace the neutro-
phil testing as described below.
I
Q.
O
cr
i-
U)
z
I
o
UJ
o
o
I
QL
o:
UI
H-
u
<
CO
100
50
80-95
74
52
60
G5
G8
G9
NO
COMPD
100
TRICHLORO
-DICHLORO-'ETHANElTETRACHLORO
50-
ETHANE |lOppm
lOppm
65
65
ETHYLENE
lOppm
55
Figure 43.
Effects of sewage fraction
concentrates and known sub-
stances on phagocytosis
using neutrophils.
Monocyte testing—
The Ficol Hypaque method was used to separate mononuclear
cell populations from peripheral blood (17). Approximately 20-
40% of the cells were monocytes, the rest being lymphocytes.
After exposure to known toxic substances or concentrated water
fractions for 30 minutes at 25°C the suspension was used as
effector cells to attack chromium51-labeled red cell targets
90
-------
which were coated with a specific antibody. The monocytes and
lymphocytes recognize the Fc portion of the immunoglobulin mole-
cules on the target cell and lysis is determined by release of
the radioactive chromium label.
The results for monocyte testing are reported for both
known substances and concentrated water fractions. Hexanal
and tetrachloroethylene gave graded responses (Figure 44). The
remainder of the compounds showed virtually no toxicity at 0.1
and 1.0 ppm but all had profound effects at 10 ppm. These sub-
stances included the following: chloroform, dimethyl disulfide,
trichloroethylene, n-octane, m-xylene, p-xylene, toluene, and
carbon tetrachloride. With regard to concentrated wastewater
fractions from Pomona, toxicity generally increased as G number
increased (Figure 45) in accordance with the results obtained
on neutrophils (Figure 43) and the data on ATP pools (Figure 39).
Carbon treatment (G4a) greatly reduced the toxicity of G4, con-
firming the results of the metabolic determinations.
DISCUSSION
The first year of this grant began with the development of
the human neutrophil system as. an assay system for water toxic-
ity. Initial results were encouraging in that toxicity could be
detected with the human neutrophil assay. However, experience
over the next year led to two major difficulties:
1) Correlation of the human neutrophil bactericidal activi-
ty with human neutrophil metabolic activity was poor
85% of the time. This was predominantly due to the
insensitivity of the human neutrophil bactericidal
system.
2) The number of samples that could be screened at one
time using the human neutrophil system was limited.
Only three or four samples could be assayed in one day.
Although a considerable number of assays of toxic
chemicals and water concentrates were performed using
the human neutrophil assay system, this limitation
meant that screening of large numbers of samples was
impossible.
The human ADCC system answered both these problems. Reli-
ability of the human ADCC assay has proven outstanding. It was
after one year's experience with the ADCC assay that we were
convinced of its reliability and sought to apply this assay to
the water reuse project. In our hands the intra-assay variabil-
ity of ADCC is 5%. Internal controls are available with each
determination of the assay to assure its interpretability.
Included are positive controls (cells, antibody, and target) and
negative controls (cells, and target, antibody and target, and
target alone). These controls have proven to be consistent.
91
-------
00
£ 50
UJ
_j
LU
0
100-
CHLOROFORM
50
0
100
TETRACHLORO-
ETHYLENE
50
0
n-HEXANAL
I
O.I 1.0 lOppm O.I 1.0 lOppm O.I 1.0 IDppm
\t>
O "00
EC
I
O
50
0
100-
n-OCTANE
m-XYLENE
100-
p-XYLENE
0
O.I 1.0 lOppm O.I 1.0 10 ppm O.I 1.0 lOppm
Figure 44. ADCC assay: killing power as a function of
concentration of known toxic substances
using monocytes. (continued)
92
-------
100
0
LJ 50
UJ
_J
UJ
cr
0
00
DIMETHYL
DISULFIDE
50
TRICHLORO
ETHYLENE
5
D
0
O.I 1.0 10 ppm O.I 1.0 lOppm
TOLUENE
100
50
CARBON TETRA
CHLORIDE
0
O.I 1.0 10 ppm O.I 1.0 10 ppm
Figure 44 (continued)„ ADCC assay: killing
power as a function of concentration
of known toxic substances using
monocytes.
93
-------
u
g 100
Ul
LJ
oc
5 80
g 60
I
O
u.
0 40
z
o
SJ 20
O
UJ
or
>S n
-
_
-
-
tf* \j
15
Gl
r^i
G2
15
G3
87
98 95
CARBON
TREATED
G4
g
\j
n
G4 G4A
24
G5
78
G6 G7 FLUSH
PEAK
Figure 45. Comparative toxicities of fractions from
Pomona, California wastewater by chromium
release.
94
-------
CONCLUSION
The conclusions from these studies indicate:
1) Neutrophil phagocytosis and killing power are not gen-
erally suitable for large scale toxicity studies of the
type needed for on-line water monitoring.
2) Monocytes are more stable than neutrophils and large
batches can be processed at one time. The depression
in monocyte function induced by toxic compounds is,
with few exceptions, most significantly observable in
the concentration range 10 ppm and corresponds to
metabolic changes observed. Relatively few compounds
have any great effect on the monocyte function at the
0.1 and 1 ppm concentrations.
3)' A wide range of responses was easily elicited by con-
centrated wastewater fractions and wastewater treatment
steps were clearly detected by this assay.
95
-------
REFERENCES
1. Bellar, T.A. and J.J. Lichtenberg. Determining Volatile
Organics at Microgram-per-liter Levels by Gas Chromato-
graphy. J. Amer. Water Works Assoc., (56>:739, 1974.
2. Bertsch, W., E. Anderson and G. Holzer. Trace Analysis of
Organic Volatiles in Water by Gas Chromatography/Mass
Spectrometry with Glass Capillary Columns. J. Chromatogr.,
112:701, 1975.
3. Thomason, M., M. Shouts, W. Bertsch and G. Holzer. Study
of Water Treatment Effects on Organic Volatiles in Drinking
Water. J. Chromatogr., 158:437, 1978.
4. Grob, K. Organic Substances in Potable Water and its
Precursor. J. Chromatogr., 84:255, 1973. ibid., 90:303,
1974.
5. May, W.E., S.N. Chesler, S.P. Cram, B.H. Gump, H.S. Hertz,
D.P. Enangonio and S.M. Dyszel. Chromatographic Analysis
of Hydrocarbons in Marine Sediments and Sea Water. J.
Chromatogr. Sci., 13^:535, 1975.
6. Eiceman, G.A. Dynamic Headspace Enrichment in Trace
Volatile Organic Analysis of Aqueous Environmental Samples.
Ph.D. Thesis> University of Colorado, Boulder, Colorado,
1978.
7. Little, J.N. and G.J. Fallick. New Considerations in
Detector-Application Relations. J. Chromatogr., 112:389,
1975.
8. Creed, C.G. Liquid Chromatography Simplifies Isolating
Organics from Water. Research/Development, 40 September,
1976-
9. Smillie, R.D., A.A. Nicholson, 0. Meresz, W.K. Dubolke,
G.A.V. Rees, K. Roberts and C. Fung. "Organics in Ontario
Drinking Waters, Part III". Ontario Ministry of the
Environment, Rexdale, Ontario, 1977.
96
-------
10. Sievers, R.E., R.M. Barkley, G.A. Eicemen, R.H. Shapiro,
H.F. Walton, K.J. Kolonko and L.R. Field. Environmental
Trace Analysis of Organics in Water by Glass Capillary
Column Chromatography and Ancillary Techniques. J.
Chromatogr., 14J2:745, 1977.
11. Katz, S., W.W. Pitt, C.D. Scott and A.A. Rosen. The
Determination of Stable Organic Compounds in Waste Effluents
at Microgram-per-liter Levels by Automatic High-Resolution
Ion-Exchange Chromatography. Water Res., 6_:1029, 1972.
12. Tardiff, R.G. and M. Deinzer. Toxicity of Organic Compounds
in Drinking Water. EPA Water Supply Lab, Cincinnati, Ohio.
Presented at the 15th Water Quality Conference, University
of Illinois, 1973.
13. Holmren, H. and H.J. Day. Adenine Nucleotides and Platelet
Function. Ser. Haemat., 4^:28, 1971.
14. Atkinson, D.A. Cellular Energy Metabolism and its
Regulations. Academic Press, 1977.
15. Solomons, C.C. and N. McDermott. Use of High Performance
and Thin Layer Chromatography in the Rapid Detection of
Human Cellular Toxicity of Environmental Substances.
Presented at the 12th Annual Conference on Trace Substances
in Environmental Health, University of Missouri, Columbia,
Missouri, June 5-8, 1978.
16. Poplack, et al. Blood, 4_6_:6, 1976.
17. Weston, W.L., R.D. Dustin and S.K. Hecht. Quantitative
Assays of Human Monocyte-macrophage Function. J. Immunol.
Methods, 8_:213, 1975.
97
-------
APPENDIX A
TESTING OF WASTEWATER FRACTIONS FOR THE PRESENCE OF
POSSIBLE CARCINOGENIC SUBSTANCES EMPLOYING THE
AMES SALMONELLA/MAMMALIAN MICROSOME MUTAGENCICITY TEST
by
Dr. Elias Balbinder
Cancer Research Center
American Medical Center
OBJECTIVE
The objective of this work was to supplement the toxicity
testing of Dr. Solomons with a test for mutagenicity, While the
time required for the Ames test precludes it from being part of
an on-line, fast response system, there is a need to test for
substances which might be carcinogenic. Thus far the Ames test
is the fastest, least expensive, and most reliable in vitro test
that gives information relating to carcinogenicity.Since there
are indications of the existence of mutagens in the effluent
from the Denver Metro Sewage Plant, it is of particular impor-
tance to monitor the effectiveness of the advanced treatment
system in removing these mutagens. Concentrations and fraction-
ation of the samples was carried out by Dr. Walton, and then the
various fractions were tested by us, using the Ames Salmonella
test, as described below.
GENERAL DESCRIPTION
The rational behind the Ames test is based on two major
observations. First, about 90% of all carcinogens tested thus
far are also mutagens (1). Second, mutagens react with and alter
DNA, and DNA has the same double helical structure and the same
four nucleotides in all living beings. Thus, it is quite reason-
able to use bacteria (or any living organism) for the detection
of potential mutagens for humans. Bacteria are easy to culture
and the availability of well-characterized mutant strains make
them a very attractive system to use for that purpose.
Ames and his collaborators have developed'several bacterial
tester strains containing different types of histidine mutations
(2) .
98
-------
One strain (TA 1535) can be used to detect mutagens causing
base-pair substitutions and two (TA 1537 and TA 1538) can be used
to detect various kinds of frameshift mutagens. In addition to
the histidine mutation, each tester strain contains two addition-
al mutations that greatly increase its sensitivity to mutagens;
one causes loss of the excision repair system and the other loss
of the lipopolysaccharise barrier that coats the surface of the
bacteria. The sensitivity of the test has been increased by
introducing a resistant transfer factor (R factor) carrying a
gene for resistance to ampicillin into the strains just
described. This has made it possible to detect classes of car-
cinogens not previously detected (2).
The test can be performed in several ways, each with its
advantages and limitations. The standard procedure is the
plate incorporation assay. In this procedure an overnight
culture of a bacterial tester strain is mixed with the sample
to be tested (presumed mutagen) in molten agar at 45°C, and the
mixture is poured on top of a plate containing minimal glucose
agar. Only histidine-independent revertants can grow on this
medium, and by counting the number of revertants a measure of the
strength of the mutagen can be determined. Revertants are gen-
erally scored after two days of incubation. Many compounds are
not mutagenic per se, but can give rise to mutagenic substances
if acted upon by enzymes in a mammalian host (metabolic activa-
tion) . To test for this possibility, a fraction of rat liver
homogenate (S-9 fraction) can be added to the above mixture.
This is the standard method that has been used for validating the
test using hundreds of chemicals. For initial screening of a
chemical testing, concentrations over a wide dose range (say 0,2,
20 and 500 yg per plate) are recommended both in the presence
and absence of the standard S-9 mix. A positive or questionable
result should then be confirmed by demonstrating a dose response
effect using a narrower range of concentrations. In general,
Ames and co-workers find that for most mutagens a concentration
range exists when there is a linear dose response relation, and
the revertants per plate reported for any mutagen should be taken
from this region of the curve. In each experiment positive
mutagenesis controls are routinely included using diagnostic
mutagens to confirm the reversion properties of each strain.
A very useful variation is the use of spot tests. Spot
tests are the simplest way to test compounds for mutagenicity and
are particularly adaptable for the initial rapid screening of
large numbers of compounds in a short period of time. There are
several advantages to the spot test and it is often useful to
test.all new compounds by this method before doing the standard
plate incorporation test. No solutions are necessary since a
few crystals (or y& of liquid) can be put directly on the agar
surface; also, since the compound diffuses out from the central
spot, a range of concentrations are tested simultaneously. The
spot test affords a preliminary indication of the toxicity of the
99
-------
chemical for the bacteria by the size of the zone of inhibition
of the background lawn of bacterial growth around the spot; it
further shows whether or not the S-9 mix is necessary for
mutagenicity, and in the case of a positive result, indicates
which tester strain should be used for the dose response curve.
The spot test is primarily a qualitative test, and although very
useful, has distinct limitations. It can only be used for the
detection of chemicals which are diffusable in the agar, and thus
most polycyclic hydrocarbons and other water insoluble chemicals
are not easily detected by this procedure. It is also much less
sensitive than the standard plate test as only relatively few
bacteria on the plate are exposed to the chemical at any parti-
cular dose level.
By the use of this test about 300 carcinogens and non-
carcinogens have been tested for mutagenicity (1) and a high
correlation between carcinogenicity and mutagenicity demon-
strated; ninety percent (156/174) of carcinogens are also muta-
gens while few non-carcinogens show any degree of mutagenicity.
The carcinogens comprise a wide variety of chemical types in-
cluding alkylating agents, nitrosamines, polycyclic hydrocarbons,
fungal toxins, aromatic amines, nitrofuran carcinogens, a variety
of neoplastic agents, and antibiotic carcinogens such as
adriamycin, daunomycin, and mitomycin C. Also, known human
chemical carcinogens which have been tested are positive. These
include B-naphthylamine, benzidine, cigarette smoke condensates,
bis-chloromethylether, aflatoxin B-, , vinyl chloride, 4-amino-
biphenyl, etc (1).
The Ames test can be used to test not only pure compounds
but complex mixtures as well, and thus has a very wide range of
applications. Because linear dose response curves are usually
observed the test is quite valuable as a bio-assay in identifying
and purifying mutagenic components in complex mixtures. For
example, it has been useful in determining the mutagenic activity
of cigarette smoke condensates and 12 standard smoke condensate
fractions (3), in commercial hair dyes (4), and soot from city
air (mentioned in 1). Other applications of the salmonella
mutagenesis test include the detection of mutagenic metabolites
in urine (2). This would be extremely useful in analyses of
human urine from individuals who may be heavily exposed to pre-
sumptive carcinogens.
In summary, the Ames test is ideally suited for rapid test-
ing of environmental pollutants to determine their mutagenic and
carcinogenic potential. Positive results in this test would
clearly indicate that the chemical (or chemicals) in question
represent a potential human health hazard and should be thorough-
ly tested in animal systems; where extensive human exposure has
occurred, appropriate epidemiological studies should be perform-
ed.
100
-------
To test compounds on unknown mutagenicity we use the
following protocol:
1) Spot tests on all tester strains using the highest
possible concentrations of compound or mixture being
tested.
2) If results are negative, perform full plate tests
(plate incorporation assays) on all tester strains
again using the highest possible concentration of com-
pound. This is to see whether the compound has weak
mutagenic activity and this is difficult to ascertain
in a spot test.
3) If results of initial spot tests are positive, a dose
response study employing the tester strain giving the
best response will be carried out.
We tested nine fractions of Denver wastewater (samples Gl
through G9) supplied by Dr. Walton. Approximately eight liters
of secondary sewage effluent were preconcentrated by fractional
freezing, then processed in the manner outlined in the text,
ending with 3 m£ of each of the fractions. Full plate tests
were conducted in duplicate with each strain, both with and with-
out S9 (microsomal rat liver fraction). In each case we added
0.3 m£ of each sample (undiluted) per plate. The following
controls were routinely carried:
1) Checked tester strains for normal response to mutagens
by using a standard set of known chemical mutagenic
agents such as nitrosoguanidine, 2-aminofluorene and
daunomycin.
2) Possible effect of methanol (which was the solvent
employed) on the observed results was tested by running
a set of control plates to which methanol was added to
the bacteria.
3) Standard set of control plates containing only bacteria
(with and without S9) was used to determine spontaneous
reversion frequency.
The results of these tests are shown in Table A-l, which
gives the average number of revertants per plate. According to
the guidelines set up by Ames and his collaborators, an increase
in the number of induced revertants less than two-fold over the
spontaneous numb.er was not significant. The results of these
tests clearly indicated that mutagenic substances were present
in samples Gl, G2, and G3. While we found no indication of
mutagenic activity in samples G4 through G9, this should not be
interpreted to mean that these samples were free of potential
mutagens. It is possible that mutagens were present, but that
their concentrations were too low to be detected by our assay.
101
-------
TABLE A-l. RESULTS OF FULL PLATE AMES TESTS
ON DENVER WASTEWATER FRACTIONS
Strain
S9
Control
Gl
G2
G3
TA
TA
TA
TA
TA
1538
98
1537
1535
100
+ 34
31
+ 72
28
+ 16
15
+ 33
18
+ 48
15
299
46
450
15
100
106
742
35
285
30
162
60
108
101 (?)
475
138
run out
run out
run out
run out
371
32
238
90
171
91
run out
run out
run out
run out
The mutagenic activity displayed by the first set of samples
was important enough to warrant further confirmatory tests. Con-
sequently, Dr. Walton processed more secondary effluent, taking
8 liters each of both chlorinated and unchlorinated wastewater.
As before, each sample was preconcentrated by freezing and
fractionation, and we were supplied with 12 concentrated frac-
tions of chlorinated and unchlorinated effluent. This time we
used .1 and .2 mi of each sample.
As shown by Table A-2, there was no positive mutagenic
response; values for both chlorinated and unchlorinated samples
are within an acceptable range of the controls.
CONCLUSIONS
The difference between the mutagenic activity of the two
sets of samples is most probably due to the difference in the
times of collection of the effluent. This would indicate that
the concentrations of compounds and mutagens, as well as the
types of compounds, vary underlining the necessity for routine
mutagenicity testing.
102
-------
TABLE A-2. RESULTS OF CONFIRMATORY AMES TESTS ON
DENVER WASTEWATER FRACTIONS
Strain
TA 1538
TA 98
TA 1537
TA 1535
TA 100
S9 Sample
+ chlorinated
unchlorinated
chlorinated
unchlorinated
+ chlorinated
unchlorinated
chlorinated
unchlorinated
+ chlorinated
unchlorinated
chlorinated
unchlorinated
+ chlorinated
unchlorinated
chlorinated
unchlorinated
+ chlorinated
unchlorinated
chlorinated
unchlorinated
Control
25
12
15
4
65
37
34
8
10
12
13
19
29
20
28
22
17
24
30
28
Gl
44
39
12
40
36
78
53
42
10
34
21
23
28
15
26
26
16
21
31
27
G2
26
33
28
18
62
68
52
51
14
32
5
17
26
11
29
22
30
20
31
20
G3
30
55
28
13
56
70
61
56
10
26
15
33
35
21
23
20
23
22
25
22
REFERENCES
1. McCann, J., E. Choi, E. Yamasaki, and B.N. Ames.
B.N.-P.N.A.S. 7^:5135-5139, 1975.
2. Ames, B.N., J. McCann and E. Yamasaki. Mutation
Research 3^:347-364, 1975.
3. Kier, L.D., E. Yamasaki and B.N. Ames. B.N.-P.N.A.S,
71:4159-4163, 1974.
4. Ames, B.N., H.O. Kammen and E. Yamasaki. P.N.A.S. 72
2423-2427, 1975.
103
-------
APPENDIX B
USE OF HIGH PERFORMANCE LIQUID CHROMATOGRAPHY AND THIN LAYER
CHROMATOGRAPHY IN THE RAPID DETECTION OF HUMAN CELLULAR
TOXICITY OF ENVIRONMENTAL SUBSTANCES
by
Clive C. Solomons, Ph.D. and Nancy McDermott, B.S.*
University of Colorado Medical Center, Denver, CO. 80262
ABSTRACT
The amount of ATP (adenosine Triphosphate) and the rate at
which it can be synthesized is important in maintaining adequate
cellular functioning. Consequently, any interference in the
storage and production of purine nucleotides can be regarded as
potentially toxic.
A series of pure organic compounds and fractions isolated
from treated sewage effluent and water treatment plant was
placed in contact with human neutrophils and platelets. Trace
amounts of U-14-C adenine were added to the cell suspensions
which were incubated at 37°C for 30 minutes. Perchloric acid
extracts were analyzed by high performance liquid chromatography
(HPLC) and thin layer chromatography (TLC) for ATP, ADP, AMP,
uric acid, hypoxanthine, and xanthine. Dilutions of the com-
pounds were used to obtain dose response curves. The HPLC
methodology is capable of automation and could perform a 5-point
dose response curve every hour. Factors involved in the
utilization of these techniques for the on-line surveillance of
renovated water for human consumption are discussed.
INTRODUCTION
Stress may be regarded as any influence which causes the
cell to increase the work it must do to maintain itself and
carry out its function. Stress effects are largely reversible
when held within physiological limits. However, significant
increases in cellular entropy content due to interference with
*Presented at the 12th Annual Conference on Trace Substances in
Environmental Health, June 5-8, 1978, University of Missouri,
Columbia, Mo.
104
-------
information processing or membrane stability within the cell can
lead to permanent pathological damage expressed in a variety of
ways depending upon which structures and metabolic pathways are
specifically affected. Serious stress effects include mutation,
transformation to malignant states, loss of energetic capability
for ion pumping, and cell death followed by lysis, which can
yield further toxic breakdown products.
The ultimate source of energy for all cellular functioning
is the ATP (adenosine triphosphate) molecule synthesized by
oxidative enzymatic catalysis involving subcellular mitochon--
drial membranes and cytoplasmic anaerobic glycolsis. ATP de-
rives its energy ultimately from the sun via edible fuels such
as carbohydrates and protein and is a universally acceptable
mediator of all cell work.
Consequently, a disturbed ATP turnover of concentration is
a serious sign of impending damage to the cell. Degradation of
ATP can form several products including hypoxanthine (HYPX),
which is salvable to some extent to reform ATP (Figure B-l).
However, the appearance of large amounts of HYPX together with
a reduced pool of ATP is a sign of severe cellular stress.
The above rationale has been used to develop toxicity de-
tection criteria for the testing of a wide variety of compounds
in drinking water, and is capable of automation and on-line
operation. A series of cells or subcellular organelles from
animal, plant or other biological origin can be used to evaluate
the environmental impact of a trace contaminant. Water can be
used directly, after adjustment of osmotic pressure, or the
water can be concentrated and fractionated and each fraction
separately evaluated. A 5-point dose response curve can be
constructed every two hours and the results evaluated by com-
puter to signal an alarm according to present criteria. The re-
mainder of this paper discusses the methodology and results
obtained using pure compounds and concentrated water fractions.
MATERIALS AND METHODS
Cells
Neutrophils and monocytes supplied by Dr. Weston and plate-
let-rich-plasma (1) which we prepared by standard methods were
presented for analysis after exposure to four or five concentra-
tion levels in the range 0-10 ppm of the test compound. Frac-
tions obtained by Dr. Walton from sewage and water treatment
plants were also tested. Five microliters of U-14-C adenine
(60,000 cmp) were added to a 0.5-milliliter suspension of cells
5000-10,000 cu mm. After incubation at 37°C for 30 minutes the
cells were cooled in ice and centrifuged at 3000 rpm for 3 min-
utes. The supernatant was removed and 0.2 m£ of cold 13% per-
chloric acid (PCA) was added to the cell pellet. After mixing
105
-------
SALVAGE
CELL MEMBRANE
HYPOXANTHINE -
XANTH.
Figure B-l. Effect of stress on ATP
metabolism.
106
-------
and centrifuging, the PCA supernatant was applied to the chroma-
tographic systems described below.
Chromatography
High Performance Liquid Chromatography (HPLC)—
A Waters & Associates HPLC apparatus with Cj8~Bondapak re-
versed phase column was used. Fifteen microliters of PCA ex-
tract was injected and 0.1M KH-PO. was pumped through the column
at 1.5 m£/min. The nucleotides and related compounds were sepa-
rated in 12-15 minutes and were detected by UV absorption at
254 nm. Identification was obtained using retention time and
the addition of known compounds to the test PCA supernatant. The
effluent from the HPLC system can be pumped through a flow-
through 'scintillation spectrometer so that the specific activity
of each compound could be computed. In these experiments a
suitable flow-through spectrometer was not available; the
effluent was collected into twenty fractions and counted in a
scintillation spectrometer. Recovery and reproducibility of the
standards was 95-103%. The peak height was found to be linearly
proportioned to concentration, regardless of peak width, in the
range of 10~° to 10 moles of. purine compound per sample in-
jected (Figure B-2).
Thin Layer Chromatography (TLC)--
Ten microliters of the PCA supernatant were applied without
heat to a 2-centimeter strip of an Eastman Kodak Cellulose Thin
Layer Chromatogram #13255 with a mylar backing, and developed
with a mixture of water: formic acid: tertamyl alcohol: 1:2:3 for
5-6 hours (2). The chromatogram was dried in air, cut into 15,
1-centimeter strips, and placed in a 5-milliliter counting vial
with 3 m£ of scintillation fluid. The vials were counted in a
scintillation spectrometer and each count calculated as a per-
centage of the total count.
RESULTS
Adequate and rapid chromatographic separation of the purine
nucleotides and their metabolites was obtained by HPLC (Figure
B-2). The relationship between peak height and concentration
was linear throughout. This method is capable of detecting
1 x 10~ moles of purine compound in the intracellular pool.
Some raw data is shown in Figure B-3. The TLC separations were
not as sharp but did provide essential data on the rate of ATP
synthesis on the order of 10"1 moles/hr, independent of the
pool size (Figure B-4). Table B-l lists the response of neutro-
phils and platelets to various pure compounds and to fractions
obtained from drinking water. As an example, the dose response
curve for chlorofrom is plotted in Figure B-5.
107
-------
ATP
t
E
c
IO
CM
CO
Z
LU
Q
O
H
Q.
C
AMP
ADENINE
6 9
TIME(MIN)
12
Figure B-2. Elution curve of pure standards
using HPLC.
108
-------
CELLS ALONE 5/10/78
50
NEUTROPHILS-HPLC
CELLS + G3
POMONA
Figure B-3. HPLC raw data,
109
-------
TLC
n
11
i '
i
i
Plus K) ppm i
m-xylene i
10
CM
Figure B-4.
Toxic effect of m-xylene
on platelets using TLC.
110
-------
o
EH
Q
IS
CO
Q
IS
D
O
O
U
W K
f£ pq
D EH
j-i -Y*
PM fi,
15
CO
D O
O IS
H H
DH Nn
H
O Q
EH
a
co O
W PLI
i-q
W Q
EH W
H CD O 3
HO CD H IH C
OH C >i O -H
C XJ OJ X rJ -P
(0 u 3 m o c
X -H rH 4-> H O
CD JH 0 CD XJ 0
ffi EH EH 2 U —
111
-------
, .
13
U
EH
< g
U
UH -P
T r \ r*
H- ( U AH
ffi + X
(0
CN OJ
rH ft
U
0
rH
ft O
Ur i *•*.
tH O
r4 ft, ft
EH
HH
U U
0\°
O
G
O
U g
^
ft
0)
rH rH
•H -H
-P -P -P X! X!
OJ OJ OJ ft ft
H r-H rH O O
QJ QJ QJ (H ^
-P -P -P -P 4-1
CO fO fO P 3
rH rH rH QJ QJ
ft ft ft IZ JS
o^cncxj o in CT\ o ovot^un
roiHMo ro^rr^vD ro •— j CN o
CNCNJCNCN CNCsliHrH CNCN r->x>r- t-tor~
Or-lOO OrHOO OrHOO OOO
OOrHO OOrHO OOrHO OOO
rH rH rH CN
QJ
>iTS C
X -H OJ
O 0 rM
£ tO >i (0
-P X X!
QJ U O -P
g -H x! x:
1 G -P ft
CN O QJ fO
•H c G
QJ H QJ 1
G ft VD C >i
QJ O QJ X QJ QJ
rH M OH O G C
fO ft g fO X! fO fO
X! O OX! -P -P X
4-> M M 4-) CN G CN G G
H H
•H -H
X! X!
ft ft
O O
tl tl
1 1 1 >
3 p
QJ QJ
S 2
Iiii i i t i
ill i i i i
O OOO OOOO
in or^oo HOO<£>CN
CN r-HH t-^DVOI--
O O O O H O O
O OHO OOHO
O OOH OOOH
CN
QJ
C
(0
r*,
4-1 QJ
QJ G
O QJ
&H N
O G
H QJ
xi n
u o
-r) 5H
>H O
4-> H
1 Xi
H 0
H -H
H Q
(1)
3
G
•H
-P
G
O
O
112
-------
T3
CD
j3
C
•H
4J
£
0
u
• — •
rH
CQ
W
^
03
EH
rH
H
CD
U
ft
EH
rd
CH ^3
•H CD C!
X -P • 0
C tO M U to
•H H Q CD C
rH 4-1 CO O
•r) c >i S
•HO O ft
•P C n3
tO O CD C 4-1
G V-rl ^ O C
C H CD -H CD
•H SH &i 43 4-* 3
O CD O, 10 U H
c -p ^ u mm
•H to w u !H MH
U S — - 2 h W
^J ^J
CD CD
rH rH
CD CD
[ i | \
fO fO
•H rH
ft ft
o VD n o VD in
H o o m ro n
rH rH rH rH rH rH
O O O O O O
in ro in VD ^t* v*o
m in ^r in in in
o m o o o o
o CN in o o o
rH CN O O
>JD CN
rH
CD
-rH
to m
rH rH
O 3
>H W
4-1
£ rH CD
O >i W
U A O
4-> U
CD 0) 3
> e H
•H -H O
4J Q
fO ~
tP ^ -H
CD -H -H
r^ "^"^
113
-------
DOSE RESPONSE CURVE
(CHCI3)
10'10M 7oCUPool
ATP
250 +100
200 +80
150 + 60 X
Total Pool
100 + 40 C14Pool
50 + 20
0.1 1.0
ppm
10
Figure B-5. Dose response curve
for chloroform.
114
-------
DISCUSSION
The utilization of cell incubation followed by HPLC and TLC
analysis offers a considerable capability for the sensitive and
rapid detection of cellular toxicity. The system is flexible in
that it is not limited to any cell type or environmental sub-
stance, and is relatively inexpensive. The system is vulnerable
in that it needs a constant supply of normal cells or organisms.
However, continuous 5-point dose response curves can be obtained
using platelets from 10 raJl of blood per day available from a
nearby blood bank. Alternatively, one could use in vitro cell
cultures of fibroblasts or bacteria, which can be made in large
batches. As long as sufficient cells are present an accurate
cell count is not needed, provided the same suspension is used
within 5 hours for test and control. If necessary, a portion
of the cells can be diverted for functional studies or other
confirmatory investigations. The test substance is usually in
solution, but this is not an absolute requirement as the toxic
effects of exposing platelets to different surfaces can also be
determined (Figure B-6). The monitoring is capable of contin-
uous on-line operation as indicated in Figure B-7. Although
this work was concerned with ATP because of the universality of
its function in living systems,, one is not restricted to this
pathway. The same approach can be used to focus on a variety of
metabolic and enzymatic effects and any cellular function which
is of interest.
The monitoring can also be done to study internal unit pro-
cesses such as reverse osmosis permeate and reject in an ad-
vanced water treatment plant. Thus, procedures such as reverse
osmosis and activated carbon treatment, etc., can be evaluated
from the biological standpoint with reference to plant, animal,
and human cells.
Extrapolation from these results to gross human toxicity
data available in RTECS shows that certain correlations exist.
For instance, chloroform is more toxic than toluene in both the
cellular ATP response and LD50 dosages. Although there are
many factors which influence the toxic effects of a chemical,
a broad correlation is present, indicating that cellular ATP
measurements are not too far removed from clinical toxicity
effects. These measurements, in fact, offer us a more meaning-
ful method of evaluation in many cases, especially when exposure
levels are low. The results on platelets are in accordance with
previous work by other investigators on blood cells and bio-
materials (3) .
CONCLUSION
It is concluded that HPLC chromatography of cellular
purines may offer a sensitive, economic, and biologically signi-
ficant method of detecting toxicity. The system is worthy of
115
-------
ACETATE
SILICONE
RUBBER
1234
CONTACT TIME MRS.
Figure B-6.
Effect of foreign surfaces on
nucleotide metabolism. H=
irreversible loss of purines
from the platelets.
116
-------
O
z
0£
O
^
z
O
CO
O
O
u
O
CO UJ
Z> U
U
U
I—
O
3-;
to O
o<
t
<
oo
UJ
u
•
u
O
0£
Q.
<
I—
<
ex.
<
<
CO
>1
CO
tn
C
•H
H
O
-p
•H
c
i
Q)
C
•H
iH
I
C
O
T3
(U
CO
O
a
O
n
Oi
I
pa
Q)
117
-------
consideration for the automated testing of water-borne environ-
mental substances.
ACKNOWLEDGMENTS
We thank the Bonfils Blood Bank for supplying blood, and
Drs. W. Weston and Norris and R. Dustin for preparing the white
cells. Drs. Walton and G. Eiceman provided extracts of water
and sewage, and Dr. P. Predicki prepared the surfaces for
platelet studies. This work was supported by a grant for the
study of water renovation from the Environmental Protection
Agency.
REFERENCES
1. Weston, W.L., Dustin, R.D. and Hecht, S., 1975,
Quantitative Assays of Human Monocyte-Macrophage Function,
J. Immunological Methods, 8, 213-222, 1975.
2. Solomons, C.C., and Handrich, E.M., 1975 in 'Biomedical
Application of Polymers' H.P. Gregor, Ed. P.9., Plenum, New York,
3. Brash, J.L., 1977, in 'Behavior of Blood and its
Components at Interfacers' L.Vroman and E. Leonard, Eds.
Ann, N.Y., Acad. Sci., Any AA9, 283, 356, 1977.
118
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/1-79-014
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Health Effects of Consumption of Renovated Water:
Chemistry and Cytotoxicity
5. REPORT DATE
March 1979 (issuing date)
6. PERFORMING ORGANIZATION CODE~
7. AUTHOR(S)
Willard R. Chappell, Clive C. Solomons
Harold F. Walton, William L. Weston
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
University of Colorado
Environmental Trace Substances Research Program
Campus Box 215
Boulder, Colorado 80309
10. PROGRAM ELEMENT NO.
1CC614
11. CONTRACT/GRANT NO.
R803968
12. SPONSORING AGENCY NAME AND ADDRESS
Health Effects Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final 10/75-9/78
14. SPONSORING AGENCY CODE
EPA/600/10
15. SUPPLEMENTARY NOTES
16 ABSTRACT
The objective of the research has been to develop methods to separate, identify,
and measure volatile and non-volatile compounds found in secondary wastewater effluent,
and to test the suitability of the cytotoxicological assay for the substances found.
Identification and measurement of volatile organics were achieved, and known substances
were submitted,for toxicological testing. Non-volatile substances were concentrated
and fractionated and submitted for both toxicological and Ames mutagenicity testing.
Toxicity testing utilized the effect of the fractions on both metabolic and bacteri-
cidal cellular activity. The use of platelets proved to be the most suitable because
of their stability and correlation with gross human toxicity rankings. The less
polar and non-polar fractions produced toxic responses in both metabolic and bacteri-
cidal assays. An initial set of samples submitted for mutagenicity testing showed
definite activity in the more polar fractions. A second series of samples submitted
for confirmatory testing showed no activity, indicating that the concentration of
mutagens varies significantly with time.
Correspondence should be addressed to: Willard R. Chappell. Director, Environmental
Trace Substances Research Program, Campus Box 215, University of Colorado, Boulder,
Colorado 80309, (303) 492-7588
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Waste water
Potable water
Public health
Waste treatment
Organic compounds
Monitors
Water treatment
b.IDENTIFIERS/OPEN ENDEDTERMS
Waste water reuse
Drinking water
Organic analyses
Renovated water
Toxicity testing
COSATI Field/Group
68G
07C
18. DISTRIBUTION STATEMENT
Release to public
19. SECURITY CLASS (This Report/
Unclassified
21. NO. OF PAGES
131
20. SECURITY CLASS (This page)
Unc]assified
22. PRICE
EPA Form 2220-1 (Rev. 4-77)
119
-------
-------