United States
                  Environmental Protection
                  Agency
Environmental Monitoring
Systems Laboratory
Las Vegas, NV 89193-3478
                  Research and Development
 EPA/600/S4-89/035  Feb. 1990
ŁEPA        Project Summary

                  Protein Adduct-Forming
                  Chemicals for  Exposure
                  Monitoring :Chemicals
                  Selected for  Further Study
                  F. C. Schnell and T. C. Chiang
                   The present report is an expanded
                 treatment  of  those chemicals
                 recommended for further study by a
                 previous U.S. EPA internal report en-
                 titled "Protein Adduct-Forming
                 Chemicals For Exposure Monitoring:
                 Literature Summary and Recom-
                 mendations."  The  latter report
                 summarized  the literature regarding
                 adducts  formed by xenoblotlcs with
                 proteins, particularly hemoglobin and
                 serum albumin, and examined the
                 feasibility of their use as dosimeters
                 of exposure. Recommendations were
                 made  with respect  to those
                 compounds,  protein adducts and
                 detection methods best  suited  to
                 monitoring human exposure to toxic
                 chemicals,  particularly those
                 occurring at Superfund sites and
                 others of interest to the EPA.  Those
                 recommendations were summarized
                 in the form of a table  entitled
                 "Prioritized  List  of Protein Adduct-
                 forming Compounds of Interest to the
                 EPA."
                   The above mentioned, prioritized
                 list  of chemicals recommended for
                 further study is reproduced in the
                 introduction  of  the present report,
                 where the ranking scheme is also
                 explained. In subsequent sections of
                 this report,  the  individual chemicals
                 are  discussed in the same order in
                 which they appear on the prioritized
                 list. The topics covered for each
                 individual chemical are as follows:
                 manufacture and use, sources and
                 levels of exposure,  known health
effects,  metabolic detoxification
and.,activation, host factors, adduct
characterization,  rates of adduct
second order rate constants), dose-
response relationships, background
adduct levels, methods of adduct de-
tection, and research needs.
  This  Project  Summary  was
developed by EPA's  Environmental
Monitoring Systems Laboratory,  Las
Vegas, NV, to announce key findings
of the research project that is fully
documented in a separate  report of
the  same title (see Project Report
ordering information at back)

Introduction
  The Environmental Protection Agency
(EPA) is charged with protecting human
health and the environment, and it has
acted  by placing restrictions and
regulations on the use of chemicals that
have been shown to be detrimental to
human health  or to the environment.
Accurate dose measurements are critical
to the evaluation of chemical-related
health  risks and to the subsequent
development of realistic regulations to
ameliorate those risks. Accordingly, the
EPA has developed an initiative designed
to develop, refine and apply appropriate
biomarkers that  can  be  used  in
conjunction  with environmental
monitoring data to provide  a better
estimate  of exposure  and  risk for
individuals and populations. Macro-
molecular adducts formed as a result of
chemical exposure are among  the
biomarkers that the  Agency,  under  its

-------
Biomarker  Initiative,  is  currently
evaluating for use  in future human
biomonitoring studies.
   A 1987 report entitled  "Carcinogen-
DNA Adducts: Introduction,  Literature
Summary  and Recommendations"
summarized the  literature concerning
adducts formed by xenobiotics with DNA
and discussed  the feasibility of their use
as  monitoring  tools  in  exposure
assessment.  Current interest in this area
is  based on  the knowledge that  certain
chemicals, including the majority  of
carcinogens  and mutagens, are or may
be  metabolized  to  form  reactive
electrophilic species.  These  reactive
electrophilic  covalently  bind   to
nucleophilic  sites   in   cellular
macromolecules such as DNA, RNA and
various  proteins. It was  recognized that
measurements  of   protein  adducts,
particularly  those of hemoglobin  and
serum  albumin, may provide a more
quantitative measure  of exposure over
longer  periods  of time than is  generally
obtainable from  measurements of DNA
adducts.
   A previous  U.S. EPA internal report
entitled  "Protein   Adduct-Forming
Chemicals For  Exposure  Monitoring:
Literature  Summary  and Recommenda-
tions"  summarized   the  literature
regarding adducts formed by xenobiotics
with proteins, particularly hemoglobin and
serum  albumin, and  examined  the
feasibility of their use as dosimeters  of
exposure.  Recommendations  were made
with respect to  those compounds, protein
adducts and  detection methods best
suited to monitoring human exposure  to
toxic  chemicals,  particularly  those
occurring at  Superfund sites and others
of interest to  the EPA. Those recom-
mendations were summarized in the form
of a table entitled "  Prioritized  List  of
Protein  Adduct-forming  Compounds  of
Interest to the EPA".
   The selection criteria were designed to
identify the most suitable protein adducts
for use  in monitoring  human exposure to
chemicals of interest  to  the EPA. Those
criteria were  as follows:
   (1) The chemical  should be identified
   as being of interest to the U.S. EPA.
   (2)  The  chemical  should  form
   measurable  amounts of a distinct, well
   characterized protein adduct.
   (3)  The  background  levels of  that
   adduct in non-exposed populations
   should be as low as possible.
   (4)  The  adduct  must be  readily
   measurable by existing  analytical
   techniques.
   (5) The adduct must be accumulated
   and eliminated in a predictable, dose-
   related manner.
   (6)  Both  exposed  and  control
   populations must be available for a
   monitoring  study  of  the  adduct-
   forming chemical.
   (7) Current levels of human exposure
   to the adduct-forming chemical should
   be associated  with  some  potential,
   adverse health effect(s).

   The present  report is  an expanded
treatment  of  those chemicals  recom-
mended for further study  by the above-
mentioned, protein adducts summary
report. The prioritized list of chemicals
recommended for further study (Table 1)
is  reproduced in the  introduction of  the
present report, where the ranking scheme
is also explained.
   In subsequent sections of this report,
the individual chemicals are discussed in
the same order  in which they appear on
the prioritized list. The topics covered for
each individual chemical are as follows:
manufacture and use, sources and levels
of exposure,  known  health  effects,
metabolic  detoxification and activation,
host  factors,  adduct  characterization,
rates  of adduct  formation (i.e.,  second
order rate constants), dose-response
relationships, background  adduct levels,
methods  of  adduct detection,  and
research needs.

Conclusions and
Recommendations

Group / Chemicals
   Of the simple alkylating and arylating
agents, ethylene oxide(EO), propylene
oxide(PO) and styrene are recommended
as having the greatest potential in human
monitoring studies using protein adducts.
EO, PO and styrene are all produced and
used in the U.S. and rank  high on the list
of adduct-forming compounds of interest
to the U.S. EPA. Also, one of the simpler
existing  methods of  protein-adduct
analysis,  the   modified   Edman
degradation procedure, is applicable  to
all three of these compounds.
   EO has  perhaps the largest research
base of  any  protein  adduct-forming
chemical. Although large amounts of EO
are produced and used by industry in the
U.S., the highest exposures result from a
relatively  minor  use of the chemical, i.e.,
the sterilization  of hospital equipment.
Because  EO   is  distributed  almost
uniformly throughout the body and reacts
directly with tissue macromolecules,  its
protein adducts more nearly reflect the
level of DNA adducts in different tis
than any other chemical studied thus
   PO, a closely-related homologi
EO,  should exhibit similar  pharrr
kinetics. The  highest exposures tc
are likely  to  occur among  produ
workers. The background levels of P
adducts are lower than those of E
adducts, making the former the
informative  biomarkers  of exposure
a  larger range  of exposure.  Resul
studies of occupational exposure t<
and PO suggest that hemoglobin a<
levels  may be  more   sensitive to
exposure to PO than EO.
   Compared with ethylating  agents
simple arylating agent  styrene proc
adducts which exhibit  lower  backgr
levels and  a cleaner GC elution pi
Both  factors  would enhance
sensitivity  with  which  styrene  adc
could be detected  by  chromatogr;
procedures. Investigators  at Colu
University  are  currently conductii
study  of  occupational  exposur
styrene in boat-building  facilities of  N
and Connecticut. The heaviest expo;
occur during hull and deck laminatic
addition to hemoglobin  adducts,
study  proposes  to characterize
validate four  other  biomarken
exposure:  lymphocyte DNA  add
micronuclei, sister chromatid exchai
and unscheduled DNA synthesis.
   Human  exposure to ethylene o
propylene oxide and styrene can be:
monitored  by   measuring  t
hydroxyethyl-,  hydroxypropyl-
hydroxyphenylethyl-valine addi
respectively. Background levels of
valine adducts are lower than those  o
histidine and cysteine adducts, and
be measured  by  a modified Ed
degradation technique  which  elimir
the need to perform tedious and  1
consuming analyses   of total prc
hydrolysates or enzymatic digests.  U
basic  conditions  in the  presenci
pentafluorophenyl  isothiocyanate,
modified terminal amino acids (valini
hemoglobin) are selectively cleaved
the  protein.  The resulting pe
fluorophenyl-thiohydantoin (F5PTH)
vatives may then  be extracted
analyzed by negative chemical ioniz
mass spectrometry (NCIMS).
   Because styrene,  7,  8,  oxide,
major reactive metabolite of styrene,
a shorter biological half-life than ethy
oxide, it is possible that more sty
adducts are formed with  albumin  ir
liver cells than with hemoglobin in the
cell. However,  no  reports of  styr
albumin  adducts  were  found in
literature.  If styrene  oxide  does

-------
                 Table 1.  Prioritized List of Protein Adduct-Forming Compounds of Interest to the EPA


                  Group I • Simple Alkylating and Arylating Agents That Form N -Terminal Valine Adducts.
                          (1) Ethylene Oxide
                          (2) Propylene Oxide
                          (3) Sfyrene

                  Group II - Aromatic Amines That Form Hydrolyzable Cysteine Adducts.
                          (1) 4-Aminobiphenyl
                          (2) Benzidine
                          (3) MBOCA
                          (4) o-Toluidine

                  Group III - Chemicals That Form Hydrolyzable, but Less Well-Characterized Adducts.
                    (1) N-Nitrosonornicotine
                          (2) Benzo(a)pyrene_
                          (3) 1 - Nrtropyrene"

                  Group IV - Chemicals That Form Characterized, but Non-Hydrolyzable Adducts.
                    (1) Vinyl Chloride
                          (2) Ethylene Dichloride
                          (3) Acrylonitrile
                          (4) Acrylamide
                          (5) Chloroform

                  Group V - Chemicals That Form Poorly Characterized Adducts.
                          (1) Benzene
                          (2) Formaldehyde
                          (3) 2,4-Toluene Diisocyanate
                          (4) 7,12 - Dimethylbenzanthracene
                          (5) Epichlorhydrin
                          (6) Benzyl Chloride
                          (7) Pentachlorophenol
 dducts at the N-terminal  aspartate of
serum albumin, they might be detectable
by  the modified Edrnan  procedure.
However, no efforts to  detect N-terminal
aspartic acid  adducts in  serum albumin
using the  modified  Edman  procedure
have been reported in the literature.
   In a highly desirable  alternative to GC-
MS, the modified valine  derivatives
produced  by  the  modified  Edman
degradation procedure might be detected
using immunological techniques.  If an
antibody could be raised against the
pentafluorophenyl-thiohydantoin (F5PTH)
derivative of alkylated N-terminal valine, a
rapid,  inexpensive  method (e.g.,  a
competitive ELISA) might be developed
to  monitor  human exposure to ethylene
oxide, propylene oxide  and styrene on a
large  scale,   one   which   could
simultaneously be  validated by GC-MS of
the same analyte  (i.e.,  the F5PTH valine
derivative).  To  cite  a  precedent,
antibodies  have  been  developed  and
used  to detect benzo(a)pyrene tetrols
released  by acid  treatment  of
benzo(a)pyrene-modified hemoglobin.

Group // Chemicals
   Benzidine  and 4-aminobiphenyl  (4-
 BP) are  established human  bladder
carcinogens, and benzidine and MBOCA
rank high on the list of adduct-forming
chemicals of interest to the  U.S.  EPA.
The aromatic amines are also particularly
attractive chemicals from the standpoint
of  exposure  monitoring  due  to  a
remarkable  situation  that obtains  with
regard to their hemoglobin  adducts. 4-
ABP,  benzidine, MBOCA and  o-toluidine,
among others, are all metabolized by a
pathway  that leads to the formation of
large  amounts  of  an  acid-labile
sulfinamide adduct of the 0-93 cysteine
of hemoglobin. The ultimate electrophile,
a  nitroso  compound,  is formed m the
erythrocyte  by  co-oxidation  of  the N-
hydroxylamine  metabolite and  hemo-
globin. Under mildly  acidic conditions,
the  sulfinamide  bond  is  cleaved,
regenerating the free amine  which can
then  be  extracted,  derivatized  and
analyzed   by  GC-NCIMS.   The
development of antibodies to the  acid-
released material would  make possible a
rapid, inexpensive  and  extremely
sensitive  assay for  the entire class of
compounds.
   None   of the  aromatic amines
mentioned above is still made  in the U.S.,
but benzidine (and possibly 4-ABP) is a
metabolite of the widely  used, benzidine-
based dyes, while the other four all occur
in cigarette smoke and have been
measured in studies of smokers  vs non-
smokers. Also,  measurable occupational
exposure to 4-ABP may still occur in  the
form of an unwanted  side  product  of
other  amines  or  dyes.  MBOCA, a
suspected  bladder and liver carcinogen,
is still  widely used in  the polyurethane
industry,  where  an  occupationally
exposed  population should be available.
The  research findings of  investigators at
Michigan  State University  which  are
relevant to hemoglobin-aromatic diamine
adducts in  general and MBOCA adducts
in particular, are discussed in this report.
   Any future study of human exposure
to any carcinogenic  aromatic amine
should make a point of identifying  the
acetylator phenotype of the individual
subjects   because so-called  "slow
acetylators" are known to have a higher
risk of bladder cancer than so-called "fast
acetylators." This report includes a
discussion of  the  possibilities   of
simultaneously  monitoring levels of two
different protein adducts  of an aromatic
amine, i.e., one acetylated and one non-
acetylated, in an  effort to  clarify  the
relationship between acetylator status
and  effective  exposure to aromatic
amines.

-------
Group /// Chemicals
   N-nitrosonornicotine (NNN), benzo (a)
pyrene (B  (a) P), and l-nitropyrene (1-
NOP),  like  the aforementioned aromatic
amines, form Hb  adducts the cleavage
products of which may be released into
solution by treatment with dilute acid  or
base. However, these hydrolyzable  Hb
adducts, unlike those of aromatic amines,
are formed in only small amounts and are
not as well characterized (e.g., the parent
adduct of the hydrolysis product may not
be known.)
   The major tobacco alkaloid nicotine is
the precursor to the tobacco-specific
nitrosamines-NNN and 4-(methylnitros-
amino)-l-(3-pyridyl) -l-butanone (NNK).
The latter are among  the  most important
carcinogens in cigarette smoke and are
the major carcinogens  present  in so-
called  "smokeless"  tobacco.  Both
compounds may be  activated  via  a.
hydroxylation (the  major  activation
pathway) to methylating  agents or, to a
lesser  extent,   to   4-(3-pyridyl)-4-
oxobutyldiazohydroxide which  forms
unique,  bulky adducts  in  both Hb and
liver  DNA in rats.  The latter adducts are
detectable as their hydrolysis  product  4-
hydroxy-l-(3-pyridyl)-1 -butanone (HPB).
   The likely relevance  of the  HPB-
yielding adduct to carcinogenic risk, and
the absence  of confounding sources,-
make the  tobacco-specific nitrosamines
NNN and  NNK potentially useful model
compounds  for human  monitoring
studies. To date,  however, the scientific
literature contains no report of NNN/NNK-
Hb adducts  having  been  detected  in
humans. Because  methyl  adducts cannot
be easily  attributed  to any specific
exposure,  efforts to  detect  NNN/NNK
adducts in  humans should focus  on the
HPB-yielding adduct.  The  work  of
Stephen Hecht   of  the  Naylor  Dana
Institute should be followed closely  in this
regard. If  HPB-yielding  adducts are
detectable in blood samples of exposed
humans, low adduct levels and high inter-
individual variation can be anticipated.
   Benzo(a)pyrene,  a  carcinogenic
polycyclic aromatic hydrocarbon  (PAH),
is ubiquitous in  the  environment, even
though it is not commercially produced in
the U.S. The benzo(a)pyrene diol epoxide
(BPDE) metabolite  thought to be the
ultimate carcinogen binds to Hb in vivo.
Treatment of the  BPDE-modified Hb with
dilute acid  releases a minor, acid-labile
adduct into solution as the corresponding
BPDE  tetrol. The acid-released  tetrols
may then  be analyzed  using  HPLC-
fluorescence spectroscopy.
   However,  the  vast majority of B(a)P-
globin  adducts  are  relatively stable  to
acid, and some effort should be made to
characterize  these  adducts, some  of
which may possibly be better biomarkers
of B(a)P  exposure than  the  tetrol-
releasing  adducts. Intact B(a)P-Hb
adducts have been  analyzed by  laser-
induced fluorimetry  and by  competitive
ELISA.  In   both  cases,  Hb  was
enzymatically digested  prior to analysis
to expose the intact adducts and samples
were enriched in  B(a)P-modified peptides
by elution from  an  immunoaffinity col-
umn.
   The highest  level of B(a)P adducts
occurs in liver proteins, of which  newly
synthesized  albumin is one, and  total
binding  of  radiolabeled  B(a)P  is
reportedly much higher  in serum albumin
than  in Hb.  An  investigation  of  B(a)P-
serum albumin  adducts  is  therefore
recommended. Serum  albumin adducts
have  proven  very useful for monitoring
exposure  to  aflatoxin   B,,another
compound that binds poorly to Hb.
    1 -Nitropyrene, a widespread environ-
mental PAH, is  neither  produced  nor
used commercially in the U.S. Exposure
to this  carcinogenic  PAH  occurs  via
inhalation of airborne participates,  diesel
emissions, coal  fly  ash, carbon  black
photocopier toners  and smoke from
nitrate-fortified cigarettes. The formation
of a  single  major Hb adduct is linearly
related to the oral dose of 1-nitropyrene
in rats. The acid-labile Hb adduct, which
is released  into solution during the pre-
cipitation of globin in acidic acetone, has
been  tentatively identified  as the  cy-
steinyl sulfinamide of a  ring-hydroxylated
N-hydroxylamine meta-bolite. This  ad-
duct  may be useful for monitoring  human
exposure to 1-nitropyrene. However,  no
studies have  yet been  done  in humans,
and  the  adduct  needs to  be further
characterized in animals.

Group IV Chemicals
   Vinyl chloride (VC), the most important
of the industrial vinyl monomers,  is
heavily used  in the  plastics  industry  for
the manufacture of polyvinyl chloride and
copolymers.  While  the  potentially
exposed population is probably quite
large, actual levels of exposure to VC in
the workplace are probably quite low (i.e.,
below  the 1  ppm,  8 hr time  wieghted
average (TWA) standard promulgated  by
OSHA in  1983), because VC   is  an
established human  carcinogen.  In fact,
hepatic angiosarcoma, a rare form  of
neoplasia,  occurs  almost  exclusively
among VC reactor  cleaners.  A  protein
adduct-based  method for  expos
monitoring would be particularly usef
the case  of  vinyl  chloride,  since
reliable method currently exists for r
itoring exposure to VC  at concentral
below 5 ppm
   The reactive metabolites of VC
chloroethylene  oxide (CEO)  and
chloroacetaldehyde  (CAA), introduci
oxoethyl groups at the  sulfhydryl gr(
of cysteine, the 1-N and 3-N positior
histidine,  and the amino  nitrogen o
terminal valine. The acid-stable add
S-(2-oxoethyl)cysteine and N-(2-oxoe
histidine are currently  of limited use
monitoring VC exposure in humans.
N-terminal  valine  adduct should \
more potential  as  a biomarkei
exposure,but  the modified  Edi
degradation procedure  has,  appare
not yet been applied to the study of
protein  adducts,  nor  have VC-se
albumin adducts  been investigate*
might  prove  worthwhile,therefore
examine serum albumin as well as
using the  modified Edman degrade
method.
   Ethylene dichloride  (EDC), the lar
volume chlorinated organic compc
currently  produced  in  the  U.S., is i
primarily  in the  manufacture  of \
chloride. EDC and VC are metabolize
much  the same  way;  the  epo:
metabolites of  both produce  2-oxo<
adducts which, when hydrolyzed
analysis, are  reduced to  2-hydroxy*
adducts. Thus, the protein adducts of
EDC  and  EO  (which yields  2-H
adducts  directly)  cannot  easily
distinguished  from one  another  i
protein  hydrolysate. As  with VC,
modified Edman procedure has not
been  used to  assay  levels  of  th<
oxoethylvaline  adduct.  The  glutath
(GSH)  conjugate  of EDC also forrr
unique, bulky  DNA adduct, S-[2-
guanyl)ethyl]GSH, which, measured
urinary metabolite, might serve as a n
specific biomarker for EDC exposure
related compound,  ethylene dibrorr
has been shown  to  bind  to  se
albumin (139 nmol/g) in treated rats.
   Acrylonitrile  (ACN)  is  an impoi
industrial monomer used extensive!
the  manufacture  of  synthetic  fib
rubbers and  resins  for  a  variety
consumer  goods.  It  is  also usec
produce acrylamide (AN), an  indu;
compound  used in the manufactun
polymers  for  water  treatment. I
chemicals are  neurotoxic in  mami
and carcinogenic in rodents,
exposures are expected to be kept
minimum.  Animal data (the  only
available)  indicate that  both of tl

-------
direct-acting  agents  are  detoxified
primarily by conjugation with GSH,  and
'both are direct-acting agents that may
also form  protein adducts via their
epoxide metabolites. When modified Hb
is hydrolyzed for analysis, both the nitrile
group  of  acrylonitrile  and  the  amide
group of acrylamide are transformed  into
carboxyl  groups. Thus, the  adduct
formed  by direct Michael  addition of
either  compound  to cysteine  will  be
analyzed as  S-(2-carboxyethyl)cysteine
(the major adduct), while  the adduct
formed by reaction of  the epoxide
metabolite  of  either  chemical with
cysteine will be analyzed as  S-(2-
carboxy-2-hydroxyethyl) cysteine  (a
minor adduct).
   The  consequence  for  biological
monitoring  is  that acrylamide and
acrylonitrile exposures  cannot  be
distinguished from one another using the
hydrolysis  products of their cysteine
adducts. However, if both  glycidamide
and glycidonitrile form N-terminal  valine
adducts, the milder pH changes involved
in  the  modified Edman degradation
procedure might  leave the  amide and
nitrile groups of the  respective  adducts
intact. If this  approach  were successful,
then the glycidamide-and  glycidonitrile-
valine adducts might be  separated  and
analyzed  by  GC-MS  (provided  the
appropriate column were used) or by  LC-
MS. If  ACN/AN-adducts of  N-terminal
valine  cannot be detected in  Hb  of
exposed animals, then  ACN/AN-adducts
of N-terminal aspartate should be sought
in serum albumin.
   Chloroform is produced and  used in
the U.S. principally for use as a reactant
in the manufacture of freon (fluorocarbon-
22) and fluorocarbon plastics, and  as an
extractant  and industrial solvent  in  the
dye and drug  industries.  Production
workers make  up  the  best exposed
population, but  again,  exposures  are
expected to be low. Although chloroform
is not  mutagenic in bacteria, it causes
cancer in rodents, and is classified by the
U.S. EPA  as  a probable  human car-
cinogen.
   In  vivo binding of' radiolabeled
chloroform  to rat Hb increases linearly
over a  wide  range of  dose  and  is
eliminated at  a rate consistent with  the
lifespan of erythrocytes.  In vitro  the
reactive metabolite  of chloroform,
phosgene, forms  an  unusual cyclic  2-
oxothiazolidine derivative with cysteine in
Hb. This adduct  is hydrolyzed during
preparation  for  GC-MS  analysis and
measured as  N-hydroxymethyl cysteine.
tHowever, it should be noted that there is
some controversy over  the identification
of  N-hydroxymethyl  cysteine as  the
major product by GC-MS,  because the
hydroxymethylamino moiety should have
been too unstable to be isolated as such.
   In view of  the unreliability  of  blood
levels and breath levels of chloroform for
monitoring exposure,  a protein adduct-
based method would  be  particularly
useful.  However, no studies of  the
formation of chloroform adducts in human
Hb have yet appeared in the  literature.
More research is  needed to identify the
chloroform-protein adducts formed in
vivo.
Group V Chemicals
   Benzene is  produced in great quantity
(1.72 billion gallons in 1988) in the U.S.
where it is widely used  as a chemical
intermediate in the  synthetics industry.
Because it is produced by natural as well
as  manmade sources,  benzene is
ubiquitous in  the  environment. The
largest number of people are exposed to
benzene by inhaling automobile exhaust
and cigarette smoke. An  established
bone marrow  toxin, benzene has been
associated with myelogenous leukemia in
man, and the  U.S.  EPA classifies
benzene as a Class A human carcinogen.
Hence,  the Agency's  interest in  finding
improved methods  of  monitoring
exposure to this chemical is well-placed.
   However, more research  on  benzene-
protein adducts is needed  in all  areas,
from adduct  identification to method
development, before the question  of the
feasibility of using protein-adducts to
monitor human exposure to benzene can
be  adequately addressed. In particular,
efforts should  be made to  identify  and
characterize an N -terminal valine adduct
of benzene in Hb of animals and humans.
It is recommended that work on benzene-
Hb adducts at the  Lovelace Inhalation
Toxicology Research Institute be followed
closely.
   Formaldehyde is of  interest to the U.S.
EPA because  of  (1)  its high level of
production (2.81 million tons in 1985) and
widespread use (e.g.,  in  particle  board,
plywood  and   ureaformaldehyde
insulation),  (2) the  opportunity  for
exposure in the general  population (e.g.,
in mobile homes and remodeled offices),
and  especially (3) the report that high
doses cause nasal  cancer  in  rats.
However, the induction of nasal cancer in
rats by 6-15 ppm formaldehyde--4-5 ppm
is intolerable to most  humans-does not
appear to be particularly relevant  to the
human situation. In  any case, it  is not
presently   feasible  to  monitor
formaldehyde  exposure  using  protein
adducts,  because  no  stable
formaldehyde-protein adducts have been
described  in the  literature. The DNA
protein cross-links that  are formed by
formaldehyde are  either unstable  or
rapidly repaired.
   2-4-Toluene diisocyanate is  a highly
reactive  chemical   used  in  the
manufacture of  rigid  polyurethane
products.  Because  it is such  a  potent
allergen,  OSHA has set its TLV at 5 ppb
(parts per billion), 8 hr-TWA, and 20 ppb
for 10 minutes. TDI also causes cancer in
rats and female (but not male) mice when
administered  by  gavage.  Although  the
isocyanate functional groups of TDI are
reported  to react extensively with -OH, -
SH,  or -NH  groups  on  proteins, no
specific TDI-protein adducts appear  to
have been described in the  literature.
Instead,  TDI  exposure has   been
monitored with  variable success by
measuring the titer of  antibodies  to
endogeneous adducts in sera of exposed
persons.  Given the  efficiency with which
TDI apparently  reacts  with protein,  it
should not be difficult  to chemically
identify one or more TDI-protein adducts.
Of special interest  for the purpose  of
monitoring would be any adducts  formed
at the  N-terminal  amino acids of either
Hb or serum albumin.
   7,12-Dimethylbenzanthracene (DMBA)
is an extremely potent animal carcinogen
that is  often used as a positive control in
carcinogenicity assays.  DMBA is   not
produced commercially, and its only use
is as a research chemical.  Consequently,
there is no exposed population for study.
The sulfate ester of a major microsomal
metabolite  of  DMBA,  7-hydroxy-12-
methyl-benz[a]  anthracene  (HMBA),
covalently  binds to cysteine, lysine and
methionine residues in  protein -in  vitro
However,  no  in  vivo studies of DMBA-
protein adducts  were  found in the lit-
erature.
   For  epichlorhydrin,  benzyl  chloride
and pentachlorophenol, little  information
is available apart from  the fact that the
radiolabeled  compounds do  bind  to
arnino acids or protein in vitro. No in vivo
studies have been  performed, and no
specific  protein  adducts have  been
identified.

Future  Directions
   No efforts appear to  have been made
to use  the modified Edman procedure to
detect N-terminal aspartic acid adducts in
serum  albumin.  The  reactivity  of  the
amino groups of valine and aspartic acid
should  be quite similar. The real question
is whether  the  carboxyl  group  of
aspartate will compete  with  the  amino
group  for  either  the  adduct-forming

-------
species or the Edman  reagent. If  the
method could  be demonstrated to work
with serum albumin as well as with  Hb,
then the  modified Edman  procedure
might be useful in monitoring exposure to
certain  chemicals that react poorly with
Hb.
   Although  N-terminal alkylvalines and
cysteine sulfinamides  can be  detected
and quantified  by methods that eliminate
most   of  the  time-consuming steps
associated with protein analysis, thorough
extraction and derivatization  of each
blood sample is  still required. Also, GC-
MS  is not ideally  suited  to mass
screening  due to  the  expense  of  the
instrumentation and  the level of training
January 5,  1990 12:24 PM required for
itsoperation.  The obvious solution  to
these problems is the development  of
antibodies for  the  modified  valine
residues  and  the free aromatic amines
generated by the procedures described
above. Such antibodies could be used to
quantify  hydrolyzable  adducts  in the
super-natantafter the adducts had been
released  from  the  protein  and
concentrated into a smaller volume.
   The use of antibodies that could
recognize non-hydrolyzable adducts  in
situ would cut out even more steps in the
analysis.  While  some  such adducts will
occur on the surface of  the proti
molecule,  others will  be buried
hydrophobic  clefts  and  be   le
accessible to the antibody. Neverthele
it should be possible to expose most
not all,  hemoglobin adducts by treat!
the sample  with  denaturing agents
proteases  prior  to analysis.  Me
research  is required  to  determi
whether antibodies to protein adducts c
be  developed which can  be used
quantify adduct levels in  intact prote
without  severely  reducing the sensitiv
of the immunoassay.

-------
  F.C. Schnell and T.C. Chiang are with Lockheed Engineering and Sciences Co.,
        Las Vegas, NV 89119
  Charles H. Nauman is the EPA Project Officer (see below).
  The complete report, entitled "Protein Adduct-Forming Chemicals for Exposure
        Monitoring: Chemicals Selected for  Further Study," (Order No.  PB90
        1517621 AS; Cost: $23.00,  subject to change) will be available only from:
            National Technical Information Service
            5285 Port Royal Road
            Springfield, VA 22161
            Telephone: 703-487-4650
  The EPA Project Officer can be  contacted at:
            Environmental Monitoring Systems Laboratory
            U.S. Environmental Protection Agency
            Las Vegas, NV 89193-3478
United States                  Center for Environmental Research
Environmental Protection        Information
Agency                        Cincinnati OH 45268
Official Business
Penalty for Private Use $300

EPA/600/S4-89/035
         000085833   PS

-------