CURRENT AWARENESS DOCUMENT
LINEAR AND ANGULAR FUROMARINS   PSORALENS.  ANGELICINS,

                AND RELATED  COMPOUNDS
        CARCINOGENICITY AND STRUCTURE ACTIVITY
     RELATIONSHIPS.   OTHER BIOLOGICAL PROPERTIES.
       METABOLISM.   ENVIRONMENTAL SIGNIFICANCE.
                     Prepared  by

             Yin-Tak  Woo,  Ph.D., D.A.B.T.
                 David Y. Lai,  Ph.D.
     Science Applications  Internation Corporation
                 8400 Westpark Drive
               McLean,  Virginia    22102
             EPA Contract No. 68-02-3948
            SAIC Project No. 2-813-07-409
      EPA Project Officer and Scientific Editor
                Joseph  C.  Arcos,  D.Sc.
          Extradivi sional Scientific Editor
                 Mary  F.  Argus,  Ph.D.
                      June 1986

-------
5325  Linear and Angular Furocouraarins   Psoralens, AngeLicins and  Related




Compounds





     53251  Introduction





     The linear and angular furocouraarins, mostly derivatives of psoralen and




angelicin, constitute a special class of  chemical carcinogens known as  photo-




carcinogens   Furocoumarins are naturally occurring substances  found  in  five




major plant families — parsley, rue, legume, fig and orchid    In certain




plants (e g ,  celery, parsnips), they are produced in very high amounts,




perhaps as phytoalexins, in response to fungal  infection (see Section




5.3 2552)   During the last decade, furocoumarins have captured the  atten-




tion of many clinicians, moi-eculjir biologists and health protection special-




ists because of their promise as photochemotherapeutic agents for a variety of




skin diseases, as molecular probes for the study of nucleic acid structure and




DNA repair mechanisms, and the use of 5-methoxypsoralen (5-MOP  or bergapten)




as a cosmetics ingredient





     The historical development of the use of furocoumarins as  photochemo-




therapeutic agents has been extensively described by Scott et al  (1)  and




Fitzpatrick and Pathak (2)   Furocoumarin-containing plants have been  used as




herbal medicine since 1200-2000 B C   The ancient Indian sacred book,  "Atharva




Veda" and the  Buddhist Bower manuscripts both mention the treatment of




vitiligo (a color imperfection of the skin, also known as leukoderma or  "white




leprosy") with a combination of sunlight  and extracts from the  seeds of  a




plant now classified as Psoralea corylifolia L   Another important plant, Ammi




majus, a weed  found in the Nile Valley, has similarly been used in Egypt for




centuries as a cure for this socially unacceptable skin disease.  It was not




until the 1940's that El Mofty (3) showed that  the active photochemothera-
                                      400

-------
peutic principle in seeds of Ammi majus is 8-methoxypsoralen  (8-MOP,  also

known as Methoxsalen, Xanthotoxin)    In 1974, a new milestone  was  set  when

Pamsh j^t_ _al_  (4) reported that a combination of oral administration  of 8-MOP

and dermal exposure to UVA* is highly effective in the treatment of  a  vir-

tually incurable skin disease, psoriasis   This new PUVA  (acronym  for

psoralens plus UVA) therapy has stirred widespread interest throughout  the

world   Since 1974, several thousand psoriasis patients have  received  PUVA

treatment in clinical trials in Europe and the United States    On  May  20,

1982, the U S  Food and Drug Administration approved the  use  of PUVA for the

treatment of psoriasis   The effective use of psoralens in the  treatment of

vitiligo has also led to the controversial use of 5-MOP as a melanogenic agent

in several suntan lotions (5)   Oil of bergamot, a citrus extract  which con-

tains 5-MOP, was also used in perfumery in Eau de Cologne (6,  7).  Owing to

its unique capacity to intercalate, photoreact and form both  monoadducts and

crosslinks with DNA, psoralens have also proved to be a very  useful  probe  for

the study of nucleic acid structure (8-11) and DNA repair mechanism  (8, 12)


     The potential phototoxicity of furocoumarins and furocoumarin-containing

plant substances has been known for centuries   The phototoxicity  of furocou-

marins was even mentioned in a 18th century German fairy  tale  in which the

central character was described to have suffered from photodermatitis  after

ingesting figs and exposing himself to sunlight (see 1)   Many  incidents of

phototoxicity in livestock and in humans exposed to furocoumarin-containing
*0wing to their differential biological  activities,  the  UV spectrum is
 generally divided  into  three  regions    UVA, which occupies the  320-400 nm
 region and  is also known as long wave  (wavelength)  UV,  near UV  or black light
 UVB, which  occupies the 280-320 nm  region  and  is sometimes called intermediate
 UV or "sunburn" UV   UVC, which occupies the region below 280 nm and  is also
 called short wave  UV,  far UV  or germicidal UV


                                      401

-------
plant products have been reported in the literature  (see  Section




532521)   The photocarcinogenicity of 8-MOP was discovered by O'Neal  and




Griffin (13) in 1957 in a study originally designed  to explore the possible




protective effect of the compound   The interest in  the study of  photocarcino-




genicity of furocoumarins has been rekindled recently as  the concern over




potential long-term health hazard of human use of  furocoumarins as drugs and




cosmetics ingredients grows   To date, at least 11 furocoumarins  have been




tested for in vivo or in vitro carcinogenic activity   The  intense interest in




these compounds have generated many recent publications as  reflected by the




large number of recent reviews (1, 8, 9, 12, 14, 15) on the subject   Two




recent symposia were devoted to review studies of  the photobiologic, toxico-




logic and pharmacologic aspects of psoralens (16,  17)





     53252  Physicochemical Properties and Biological Effects





     532521  PHYSICAL AND CHEMICAL PROPERTIES





     The physical and chemical properties of psoralens and  related furocou-




marin compounds have been extensively discussed in a number of reviews (1, 8,




15, 18)   Synthetically, there are 12 different ways a furan ring can be fused




to a suitably substituted ooumarin derivative to form different types of furo-




coumarins   Naturally occurring furocoumarins of plant origin consist of two




types, psoralens and angelicins   In psoralens the furan  ring is  condensed




linearly to the coumarin molecule (the 3,2-position of furan to the 6,7-




position of coumarin), whereas in angelicins the furan ring is condensed




angularly (the 3,2-position of furan to the 7,8-position  of coumarin)   The




systematic, IUPAC name for the parent compound, psoralen, is 7H-furo[3,2-




g]benzopyran-7-one   Two different systems are in  use for numbering the carbon




and oxygen atoms in the psoralen molecule   In the conventional system the




original numbering sequence of both the coumarin and the  furan moieties are
                                      402

-------
                           I
                 (Conventional)                (IUPAC)

retained, but the furan positions are distinguisned by  the  prime  sign,  such  as
1", 2',         In the IUPAC system, the atoms on  the periphery of  the  mole-
cule are numbered in a clockwise direction  starting from  the  oxygen  atom  in
the furan ring   The conventional system is  still  widely  used  in  the  litera-
ture on psoralens and is the one adopted in  this section

     Close to 30 different derivatives of  psoralen and  angelicin  have been
isolated from natural sources (1, 9)   The chemical structures of the eleven
compounds which have been tested for carcinogenic  activity  are depicted in
Table LXVIII   Some physical and chemical  properties of eight of  these  com-
pounds are summarized in Table LXIX   Most naturally occuring  psoralens and
angelicins are white to cream-colored, odorless, needle-like  crystals at  room
temperature   They are practically insoluble in cold water, sparingly soluble
in ether or boiling water, and soluble in hot ethanol,  acetone, vegetable
oils, propylene glycol, benzene and chloroform   The UV absorption  spectra of
most psoralens and angelicins show one or more absorption maxima  in  three
regions   less than 225 nm, 230-270 nm and 290-330 nm (18,  23)    The  lactone
ring in furocoumarins is susceptible to alkaline hydrolysis.  However,  if the
alkaline treatment is mild, acidification will reverse  the  reaction  to  reform
furocoumarins.  Furan-2,3-dicarboxylic acid  is obtained when  either  psoralen
or angelicin derivatives are oxidized with peroxide under alkaline condi-
tions.  Treatment of an alkoxy furocouraarin, other than a methoxy derivative,
with an acetic acid-sulfuric acid mixture leads to cleavage of the ether bond
(18).
                                      403

-------
                           Table


              Structural Formulas of Psoralen and Related Compounds
         R,
                                                        cr^o
   Psoralen (R,=R2=-H)
   5-MOP or Bergapten (R^-OCHj, R2=-H)
   8-MOP or Methoxsaien (R, =-H, R2=-OCH3)
                                  CH
   Imperatorin (R,=-H, R2=-OCH2CH=C'  3

Marmesm (R=-H)

            CH3
            I
Chalepin(R=-C-CH=CH2)
3-Carbethoxypsoralen*(R^-COOC2H5,R^R3^R^-H)

4,5',8-Tnmethylpsoralen or Tnoxsalen

                     (R,=-H,R2=R3=R4=-CH3)
 u

 Synthetic compound
     Angelic in (R,=R2=R3=-H)
     4,5'-Dimethylangel!cm(R,=R^-CH3,
             Rr-H)

-------
                                                       Table LXIX
                 Physicochemical  and Photophysicochemical Properties of Psoralen and Related Compounds'

UV-absorpt ion Fluores
Compound (synonyms m p Solubility
and abbreviations) ( C) (/Jg/tnl H20) ^max £max £365 ^ ^exclt
Psoralen 163-164 37
8-Methoxypsoralen 148 23
(8-MOP, methox-
salen, Xanthotoxin)
5-Methoxypsoralen -- 5
(5-MOP, Bergapten)
4,5' ,8-Trimethyl- 235 1
psoralen (Tri-
oxsalen)
3-Carbethoxy- 153-154 13
psoralen
Angel ic in -- 20
5-Me thy I angel i<_ in 189-190
4,5'-Dimethyl- 185 8
angel it in
295
302 12,010 1,150 360


312 14,200 950 350

298 7,950 1,680 360


317 10,900 6,230 365

299 9,870
307 10,900
298 9,350 125 330

Quantum yield
;cencec Super-
Triplet Singlet oxide
Aemis stated oxygen6 anionf
0 45, 0 12 3 3, 7 6, +++ +++
460 0 06, 0 14 1, + +


460 <0 001 04-05,+ +

430 — 53,+ ++


460 0 32, 0 35 26, ++++ +

0 33 0, +++ ++
•n- -n-
425 — * +

aSummarized from data compiled by Grossweiner (19), Rodighiero and Dal1'Ac qua  (20),  Joshi  and  Pathak  (21),  Averbeck  et
 
-------
     Tne photochemistry of  psoraiens has been  extensively studied,  the  readers




are referred to  reviews by  Song  and Tapley  (8),  Pathak (25),  Rodighiero and




DaLl'Acqua (20)  and Grossweiner  (19) for details    The essential  features  of




the photochemical changes in  psoralen brought  about  by photoactivation  are




schematically summarized in Fig   16   Psoralen  in  ground  state  (Pgo) may be




photoactivated to an excited  singlet state  (see  Section 322,  Vol   I)  by  two




possible types of electronic  excitation   (a)  excitation  of  the nonbonding




electron in the  carbonyl group to antibonding  ~JT molecular orbital  ( ft*) to




produce the  n,Tf* state and  (b) excitation  of  a  IT-electron  to the  antibond-




ing TT* state to produce the  7T,7T* state   The  resulting excited  singlet



       1                                              — ft
state ( PS) is extremely reactive, short-lived  (•*• 10   sec)  and readily under-




goes spontaneous deactivation by radiative decay  resulting in fluorescence, or




undergoes intersystem crossing (electronic spin  flip)  to  the  excited triplet


       o                                 o           o

state ( Pg)   The exicted triplet state ( n,lT* or   i7 , Tf*) > although  still




highly reactive  and short-lived, has a considerably  longer lifetime  ( — 10




sec) and may allow three competing deactivating  processes to  take  place




These are (a) spontaneous radiative decay resulting  in phosphorescence  (see




Section 322 and Suppletory  Note 3, Vol  I),  (b)  type I  photochemical  reac-




tions, and (c) type II photochemical reactions   The type I  (anoxic) reactions




involve nonradiative transfer of energy from  P   to  substrates  resulting in
                                                s



direct reaction between  P0 and substrates   In biological systems,  these
                          b



reactions include C-  cycloaddition of P  to  pyrimidines,  interstrand cross-




linking,  and conjugation to ammo acid moieties of proteins    These  reactions




are believed to  play a major  role in contributing  to the  phototoxic, photo-




mutagenic and photocarcmogenic activities of  psoraiens (see  Section




53254)   The type II (photodynamic) reactions  involve indirect,  sensitized


                        o

transfer of energy from  P  to molecular oxygen  (in  triplet  state)  to produce
                                      404

-------
                  Nonradiative
                 transfer of energy
                                       photoactivation (hl>)
                                            Fluorescence decay (V
                                       intersystem crossing (electronic spin flip)
                                            Phosphorescence decay      m PJO +• hi/)
Sensitized transfer
   of energy
                    TYPE I
          Direct photochemical reactions

             a  C^ cycloaddition of P to
                pyrimidmes

             b  Coniugation to proteins
          TYPE II
Indirect photochemical reactions

 a  Production of singlet oxygen
    33             1
     s    2       so   2

 b  Production of super oxide anion
     Fig.  16.  Photoexcited  states of  psoralen and  their  photophysical  radia-

tive decays (fluorescence and  phosphorescence) and  photochemical reactions.

TWo types  of photochemical reactions are  shown:   type I (direct photochemical

reaction), anoxic  and involving nonradiative transfer of  energy, and  type II

(indirect  photochemical reaction), oxygen-dependent and involving sensitized

transfer of energy to molecular oxygen.   Abbreviations:   Pgo, psoralen  in

ground  state; 1P8, psoralen  in singlet  state; 3PS,  psoralen in triplet  state,

^02, molecular oxygen in triplet state;  102, molecular oxygen in singlet

state;  Oo' superoxide anion.   [Adapted  from M.A.  Pathak:   NCI Monog.  66,  41

(1984).]

-------
reactive oxygen species such as singlet oxygen  ( C^) ,  superoxide  anion  (02*)




and hydroxy radicals ( • OH)   Compared  to  type  I reactions,  type  II  reactions




are relatively unimportant in contributing  to  the  photobiological effects  of




psoralens   Psoralens  have high tendency  to  intercalate  into  DNA, once  inter-




calated, psoralens are less efficient  in  producing  Qj    Furthermore,




psoralens themselves can react with  CK to  form oxidation  products  thus  effec-




tively removing  C^ from the system (8)   The  relative abilities  of several




psoralens and angelicins to yield triplet state, singlet  oxygen and superoxide




anions are summarized  in Table LXIX






     5 3 2.5 2 2  BIOLOGICAL EFFECTS OTHER THAN CARCINOGEN 1C ITY





     Toxicity and Photobiological Effects   Most of the  toxicity  studies of




furocoumar in compounds focus on their  phototoxicity    Relatively  little  data




are available on their systemic toxicity.  The  acute LDcQ of  8-MOP  in rats by




i p  injection is 470 mg/kg body weight (26)    Lower i p  LD5Q values of 189




and 158 mg/kg for male and female Sprague-Dawley rats, and  about  250 mg/kg for




CD-I mice, were found using micronized suspension of 8-MOP  (27).  The corre-




sponding oral LDcQ values of micronized 8-MOP suspension  were 791,  699  and 566




mg/kg body weight (27)   In a comparative study  (route not  specified) by




Herold _et_ _al_  (28), 5-MOP was substantially less toxic than 8-MOP,  with




reported LD5Q values of 8,100, >30,000 and 9,000 mg/kg body weight  for  5-MOP




and 875, 4,400 and 505 mg/kg body weight  for 8-MOP in I
-------
     The phototoxicity of psoralen and related compounds has been extensively

studied and was the subject of several recent reviews  (l,  30,  31)    In mam-

malian skin the most striking features of acute phototoxicity  due to psoralens

are delayed erythema and pigmentation   Under some conditions, edema and

blistering is also produced   The dose-response curve  of the phototoxicity of

psoralens is usually very steep   The phototoxic effects of psoralens are

manifested by either oral or parenteral administration or  by topical applica-

tion   The action spectrum* of psoralens lies in the 340-380 nm region, the

higher end of their absorption spectrum   Unlike for many  photosensitizers,

the photosensitizing effect of psoralens is essentially independent of oxygen

(1, see also Section 532521)

     Sporadic outbreaks of psoralen-induced photosensitization in livestock

have been noted after grazing or feeding on plants (e g ,  Cymopterus watsonn,

Ammi yisnaga, Ammi rnajus) of the parsley family (32)   Incidents of dermatitis

in celery workers, lime pickers, and consumers using the perfume "Eau de

Cologne" have also been reported   In celery workers' dermatitis, it was esta-

blished that the very high amount of 8-MOP and 4,5',8-trimethylpsoralen

produced in celery in response to fungal disease ("pink rot")  is responsible

for the phototoxicity (33, 34)   The presence of 5-MOP in  lime oil and

bergamot oil (used as an ingredient in Eau de Cologne) may account for the

occasional dermatitis in the workers and consumers (1, 6)
*The action spectrum of a photosensitizing compound is the  region of wave-
 lengths capable of eliciting photobiological response from sensitized  sub-
 jects   The action spectrum of a compound is generally expected to parallel
 its absorption spectrum   However, for furocoumarm compounds, the two  spectra
 are not in good agreement.


                                     406

-------
     The strueture-phototoxicity relationships of furocoumann compounds have

been extensively studied   More than 150 furocoumarins and related compounds

have been tested for photosensitizing activity on mammalian skin (rev  in

1)   Some of these results are summarized in Table LXX and compared to their

ability to photoreact with DNA   In general, linear furocoumarins are much

more active than angular furocoraarins.   Saturation of the 4',5'-double bond

completely abolishes activity   The introduction of a bulky substituent (e g ,

carbethoxy) at one of the two critical  double bond (3,4- and 4',5'-) of

psoralen molecule yields a compound (eg ,  3-carbethoxypsoraien) with mono-

functional character, having little or  no phototoxicity.  Compounds with a

large number of substitutents having ionic  character tend to ionize in water

and are less stable and less phototoxic   Substrtutents that alter the

resonance of the furocoumarm molecule  greatly affect its activity   The

introduction of electron-donating substituents such as a methyl group in the

5'- or 8-position enhances the phototoxicity of psoralen, whereas methyl sub-

stitution at 3-, 4-, and 4'-position decreases the activity   Introduction of

electron-withdrawing substituents (e g  , -OH, -CN, -NH2) at the same position

exhibits an opposite effect

     Comparison of the phototoxicity of furocoumarins with their photobinding

activity to DNA (see Table LXX) indicates a very good correlation between

phototoxicity and ability to form interstrand cross-links   Apparently, the

formation of monoadduct is insufficient to  cause significant phototoxicity *

This hypothesis is supported by the findings of Gange and Parrish (31) that

blue light (>380 nm) photoactivates bifunctional 8-MOP to form monoadducts
*Some exceptions to this general rule exist   Certain mono functional psoralens
 such as 4,6,4'-tnmethylangelicin are phototoxic (25)
                                     407

-------
                                  Table LXX
         Phototoxicity of Furocoumann Compounds and Comparison  with
                      Their  Ability to Photobind to DNAa
Phototoxicity
Compound in mice
Psoralen ++++
5 -MOP ++
8-MOP +++
4,5' ,8-Trimethylpsoralen (TMP) ++++
4'-Aminomethyl-TMP +•»•++
3,4'-Dimethyl-8-MOP ++
5-Diethylaminobutoxypsoralen +
3-Carbethoxypsoralen -c
Angel ic in
5-Methylangel icin
4,5 '-Dimethyl angel ic in
Ability
to form
monoadduct s
strong
strong
strong
strong
strong
strong
strong
strong
weak
strong
moderate
Ab 1 1 1 1 y
to form
interstrand
crosslinks
very strong
strong
strong
very strong
very strong
low
moderate
absent
absent
absent
absent
 Adapted from P C. Joshi and M.A  Pathak [Biochera. Biophys. Res  Commun  112,
 638 (1983)]

 Tested by skin photosensitization in mice in response to topical application
 of the compound followed bv UVA exposure

cData from M.P  Mullen, M.A  Pathak, J D. West, T.J. Harrist, and F.
 Dall'Acqua [NCI Monogr. 66, 205 (1984)].

-------
with DNA, but without forming any crosslink   Mice treated with 8-MOP plus




blue light showed no sign of phototoxicity   However, if these mice were




treated with UVA 48 hours later to convert monoadducts into crosslinks, photo-




toxicity became evident





     Mutagenicity   The rautagenicity of psoralen and related compounds has




been extensively studied   The readers are referred to an excellent review by




Scott et al  (1) for studies published prior to 1976   Numerous studies have




been reported since that time, mostly in vitro assays using a variety of test




organisms such as prophages, bacteria, algae and yeasts   The following dis-




cussion focuses mainly on studies using the Ames test and on comparative




studies.  It should be emphasized that the mutagenicity of furocoumarin and




related compounds may be affected by a number of unusual factors as discussed




below





     a)  Light   As may be expected from the photochemical reactivity of




     furocoumarin compounds, the mutagenicity of these compounds may




     differ significantly when assayed in the dark (or under subdued




     light) or when assayed with exposure to light (particularly long




     wave UV or UVA)   Their mechanism of action is believed to involve




     noncovalent complexing (intercalation) with DNA for "dark" muta-




     genicity and covalent photobinding to DNA for photomutagenicity




     (see Section 53254)





     b)  Phototoxicity   A number of furocoumarin compounds cause cell




     death after photoactivation   The bifunctional furocoumarins are




     particularly phototoxic in this respect because of their cross-




     linking activity   The mutagenicity testing of these compounds is




     often complicated by their cytotoxicity (35, 36)
                                     408

-------
     c)  Induction of iysogeaic virus   Mutagenicity assays using




     bacteria as the test organism may be affected by the  induction




     (activation) of Lysogenic virus by furocoumann compounds leading




     to lysis of bacteria (37, 38)   Interpretation of bacterial muta-




     genicity data should take this factor into consideration





     d)  DNA repair proficiency of test organisms   The proficiency of




     DNA repair capability of test organisms plays a critical role in




     determining the mutagenicity of furocoumarins in Salmonella




     typhimurium (39), Escherichia coli (12, 40), Aspergillus nidulans




     (41) and Saccharomyces cerevisiae (42)   It has been  suggested that




     the repair contributes to mutation by preventing cell lethality and




     by introducing error during the unscheduled DNA repair synthesis





     Over 20 furocoumann compounds have been tested for mutagenic activity




using the Ames test   The results of these studies are summarized in Table




LXXI   When assayed in the absence of UV-photoactivation, most furocoumarins




are either inactive or are weak mutagens toward frameshift mutants   With a




few exceptions, most mutagenic compounds do not require metabolic activa-




tion   The weak, direct-acting frameshift mutagenic activity of these




compounds is consistent with the hypothesis that, by virtue of their favorable




molecular size, they produce mutations by noncovalently complexing with




(intercalating into) DNA   Among the compounds tested for  "dark" mutagenicity,




8-MOP has been by far the most extensively studied, it was found inactive by




most investigators and weakly rautagenic in two studies (44, 50)   Psoralen-8-




glycidol ether is an extremely potent muLagen (probably one of the most potent




mutagen ever synthesized), a concentration of 0 03 uraole per plate was suffi-




cient to produce positive results   It is probably the presence of a reactive




functional group (epoxide) coupled with the molecular shape and size of the







                                     409

-------
                    Table LXXI
"Dark" and Photoactivated Mutagenicity of Psoralen
      and Related Compounds in the Ames Test
                                                           I of 2
Without UV-photoact ivat ion
Compound3
Psoralen
5-MOP (Bergapten)
8-MOP (Methoxsalen)

Imperatorin
Heraclenin

Oxypeucedaninc

Psoralen-8-glycidol ether
3-Carbethoxypsoralen
4,5" ,8-Trimethylpsoralen
(Trioxsalen)
Pyndo(3,4-c)furo(3,2-g)-
coumarin
7-Methylpyr ido( 3 , 4-c ) f uro-
( 3,2-g)coumann
Marmesin
Chalepin
Angel ic in
5-Methylangelicin
4'- or 5 '-Methylangelicin
- S-9
- (43)
- (44)
- (37, 43,
46-49)
w+ (44, 50)
+ (46)
- (46)
w* (51)
- (46)
w+ (51)
+ (51)
- (44)
- (47, 48)
w+ (44)
w+ (44)
w+ (44)
+ (46)
- (46)
- (43)
w+ (43)
- (43)
With UVA-
+ S-9 photoact ivation
- (43) + (39)
+ (39,45)
- (37, 43, - (37, 47-48)
46-48)
+ (39)
+ (46)
- (46)

- (46)



- (47, 48) - (47, 48)


+ (46)
v+ (46)
- (43) f (39)
w* (43)
- (43)

-------
                            Table LXXI (continued)
                                                                     P  2  of 2
Without UV-photoactivation
Compound3
4,4'- or 5,5'- or 4,5'- or
4' ,5-Dimethylangel icin
4,5 '-Dimethyl-4 '-hydroxy-
methylangelicin
4, 5 '-Dime thy 1-4 '-ace toxy-
angelicin
4, 5 '-Dune thy 1-4 '-dime thy 1-
- S-9
- (43)
- (43)
- (43)
- (43)
With UVA-
+ S-9 photoactwation
- (43)
w+ (43) + (43)
w+ (43)
w+ (43)
  aminoethoxymethyl-
  angelicin

Esculm (6,7-Dimethoxy-
  coumarin)
- (46)
- (46)
aExcept  where otherwise  indicated,  see Table LXVIII for structural formula.
 See below for structural  formula.
C0xypeucadanin is an isomer of heraclenin with  the substituent in the
 5-position instead of the 8-position.
 Tested  in the dark or under  subdued  light.  w+ = weakly mutagenic.
              0
                   0-CH2-CH-CX
             Heraclenin (R=-CH3)
          Psoralen-8-glycidol ether (R=-H)
           Pyndo(3,4-c) furo(3,2-g)-
                coumarm (R=-H)
          7-Methylpyndo(3,4-c)furo-
           (3,2-g) coumarm (R=-CH3)

-------
psoralen molecule, favorable for intercalation, which makes it highly muta-




genic   This is shown by the fact that replacement of the  psoralen moiety with




a phenyl ring yields phenyl glycidol ether, which is 30 times less rauta-




genic   Introduction of two sterically hindering methyl groups at the terminal




epoxide group yields heraclenm, a naturally occurring compound, which is




1,000 times less mutagenic   Interestingly, imperatorin, the unepoxized parent




compound of heraclenm, is a moderately active mutagen with or without meta-




bolic activation






     The presence of both photoreactive double bonds (3,4- and 4',5'-) in




psoralen molecules does not seem to be an absolute requirement for "dark"




rautagenicity   Blockade of the 3,4-double bond by a pyridine ring yields com-




pounds (pyr idofurocoumar ins) which s-till retain some mutagenic activity




Likewise, saturation of 4',5'-double bond does not necessarily abolish "dark"




mutagenic activity   In fact, marmesin is a moderately active "dark" rauta-




gen   Uwaifo (46) postulated that the introduction of a tertiary hydroxypropyl




group at the 5'-position may activate the 3,4-double bond, however, the




mechanism is not clear.  Among the angular furocoumarins, only 5-methylangeli-




cin is mutagenic in the dark   Three derivatives of 4,5'-diraethylangelicin,




bearing a hydrophilic group at 4'-position, are all mutagenic after metabolic




activation (see Table LXXI)






     Only six furocoumarin compounds have been tested for photomutagenic




activity in the Ames test   A number of studies (see Table LXXI) found both




8-MOP and 4,5',8-trimethylpsoralen nonmutagenic in the commonly used tester




strains (TA98 and TA100)   However using a new tester strain (TA102), which is




highly sensitive to oxidative and cross-linking agents and is proficient in




UV-excision repair, Levin et al  (39) found several furocoumarins mutagenic




after photoactivation by UVA   Quantitatively, the bifunctional furocoumarms







                                      410

-------
(8-MOP, 5-MOP and psoraien) are considerably more rautagenic  than  the raonofunc-




tional furocoumarins (angelicins)   4,5'-Diraethyl-4'-hydroxyangelicin  appears




to be an unusual compound, it is mutagenic  toward TA98  after metabolic  activa-




tion in the dark but becomes mutagenic  toward TA100  and  nonmutagenic toward




TA98 after UVA-photoactivation (43)





     Besides the Ames test, furocoumarin compounds have  been assayed in  a




variety of other test systems   With  few exceptions,  remarkably similar




results have been obtained from these studies   When  assayed in the dark,




psoraien, 5-MOP, 8-MOP and angelicin  are all weak, direct-acting  frameshift




mutagens in various strains of Eschenchia  coil, whereas 4,5' ,8-tnmethyl-




psoralerj appears to be nonmutagenic (5, 50, 52-54).   There  is  some evidence




that 8-MOP and 4,5',8-trimethylpsoralen can be metabolically activated  to a




base-pair substitution mutagen in the dark  (47)   When  assayed with UV-photo-




activation using UV-excision repair-proficient test organisms, bifunctional




furocoumarin compounds are consistently more mutagenic  than monofunctional




furocouraarin compounds   In E  coli,  the reported relative order  of photo-




mutagneic potency order is   psoraien > 8-MOP > 5-MOP >  angelicin  in one study




(54), and 8-MOP > psoraien > angelicin > 4,5'-dimethylangelicin in another




(40)   In the algal system Chlamydomonas reinhardii,  both 5-MOP and 8-MOP are




equally strong photomutagens, imperatorin  is also strongly photomutagenic




after prolonged UV irradiation, whereas 5'-methylangeiicin  is  a weak photo-




mutagen (55, 56)   In the yeast (Saccharomyces cerevisiae), 4,5' ,8-tnmethyl-




psoralen is highly photomutagenic,  the relative photorautagenic potency  is




4,5' ,8-tnmethylpsoralen > 5-MOP >  8-MOP > 4,5 '-dimethylangel icin > 3-carbeth-




oxypsoralen in the induction of nuclear mutation, and 4,5'8-trimethylpsoralen




> 5-MOP > 3-carbethoxypsoralen > 8-MOP = 4,5'-dimethylangelicin in the  induc-




tion of cytoplasmic petite mutation (57)
                                      411

-------
     Teratogemc ity   In contrast to the extensive mutagenicity data,  there  is




little or no information available on the teratogenicity of  furocoumarin com-




pounds   Herold _et_ _a_l_  (28) reported that 5-MOP exhibits no  teratogenic




effects in rabbits and rats, the details of the study are not  available




Colombo et_ al_  (58, 59) treated sea urchin sperms with psoralen and  long wave-




length UV, just before fertilizing untreated eggs, and observed abnormal




effects (formation of giant cells, polynucleated cells, impairment of




cleavage, shrinkage of cytoplasm, irregular blastomeres, and blastulas and




cytolysis of embryo) in various subsequent developmental stages   Neither




psoralen nor long wavelength UV alone exhibited any significant effect on the




development of sea urchin eggs





     53253  Carcinogenicity and Structure-Activity Relationships





     The careinogenicity of psoralen and related furocoumarin  compounds was




first studied by O'Neal and Griffin (13) in 1957 using 8-MOP   The interest  in




furocoumarin compounds was renewed recently because of their therapeutic value




in the treatment of psoriasis and other skin disorders and the commercial use




of 5-MOP in suntan lotion   To date, 11 furocoumarin compounds have  been




tested for in vivo or in vitro carcinogenic activity   Early studies led to




the conclusion that bifunctional furocoumarins (such as 8-MOP, 5-MOP)  are more




carcinogenic than mono functional furocoumarins (such as angelicin, 3-carbo-




ethoxypsoralen).  However, more recent studies using derivatives of  angelicin




indicate that the above conclusion is not necessarily true for all cases, some




raonofunctlonal furocoumarins can be more carcinogenic than 8-MOP   Like their




photomutagenicity, the careinogenicity of furocoumarin compounds can be




affected by a variety of factors as exemplified by the results of studies on




8-MOP.
                                      412

-------
     532531  CARCINOGENICITY OF 8-METHOXYPSORALEN (8-MOP)





     8-Methoxypsoralen (Methoxsalen, Xanthotoxin) was first  found to be photo-




carcinogenic in 1957 by O'Neal and Griffin (13)  in mice by i p   injection,  in




a study intended to explore the potential of the compound as a protective




agent against ultraviolet 1ight-induced skin careinogenesis   Subsequently,




the compound was extensively studied in various  strains of mice, particularly




in the past several years, because of its use in combination with UVA in the




"PUVA" treatment of psoriasis   The careinogenicity studies of 8-MOP reported




prior to 1980 were reviewed by the International Agency for Research on Cancer




(15)   The results of these and a number of more recent studies  are summarized




in Table LXXII   At the time of this writing, 8-MOP was also being tested




under the U.S  National Toxicology Program   The (.arc inogenicity of 8-MOP  is




affected by a variety of factors such as (a) the route of administration,




(b)  the type, the intensity and the timing of irradiation used to activate  the




compounds, and (c) the strain of the animals used.





     The route of administration has an important modulating effect on the




photocaremogenicity of 8-MOP   Topical application and i p  injection are




substantially more effective for the induction of skin tumors than oral




administration   By oral administration, 8-MOP has no significant carcinogenic




effect in mice of various strains as observed in several independent studies




(13, 60-64).  In fact, in one of these studies (13), 8-MOP appeared actually




to protect against far-UV-induced skin carcinogenesis (25% incidence in 8-MOP




plus far-UV-exposed mice vs. 62% in control mice exposed to far-UV alone)




The doses administered in the above-cited studies ranged from 0  6 to 40 rag/kg




body weight by gavage or 200 to 1,000 ppm in the diet   Various  forms of




irradiation (far-UV, near-UV, fluorescent light) were used to photoactivate




8-MOP, usually one hour after dosing   The duration of these studies ranged







                                      413

-------
                                 Table LXXII
 Photocarcinogenicity of 8-Methoxypsoralen  (8-MOP)  in Mice  of  Various  Strains3
Strain
Various strains
Swiss albino
Swiss albino
ICR Swiss
Swiss
C3H"/HeN
ICR Swiss
SKH-hairless-1

SKH-hairless-1
or HRS/J/Anl
Albino hairless
XVIInc/z
Swiss albino
Route
oral
oral
i p.
1 P
i.p.
i.p.
topical to ear
skin painting

skin painting
skin painting
topical to ear
skin painting
Princ ipal
organs
Irradiation affected
UV (250-400 run)
UVA
UV (250-320 mn,
320-400 mn)
UV (280-360 nm)
UV (365 nm)
UVA
UV (280-360 nm)
SSRd O290 nm)
UVA
UV (various
sources)
SSRd
UV (365 nm)
UV (300-400 nm)
Noneb
Skinc
Skinc
Skinc
Skin
Skin
Skinc
Skin
Skin
Skin
Skin
Skinc
Skin
References
(13, 60-64)
(60, 65)
(13, 60)
(66)
(67)
(68)
(66)
(69-71)
(72)
(73, 74)
(75)
(67, 76)
(77, 78)
aln all these studies, 8-MOP was found noncarc inogenic without photoactwa-
 tion.  The tumors were primarily squamous cell or basal cell carcinomas
 The duration of these studies ranged from 4 to 12 months.
cEar and/or eye region.
 Simulated solar radiation, spectrum includes UVA, UVB, visible and IR light

-------
from 4 to 12 months, although the duration was relatively short, it should be


noted that in studies of similar length, but using topical or i p   route,


8-MOP did prove to be carcinogenic   A weak photocarcinogenic effect (skin


tumor incidences of 20-35% vs  0% for control) was observed  in one  study (60,


65) in which Swiss albino mice were given 500 ppm 8-MOP in the diet and


exposed to UVA 10 minutes to 2 hours a day, 5 days a week for 6 weeks to 3


months   In contrast to its weak careinogenicity or lack thereof by oral


administration, 8-MOP was consistently found to be photocarcinogenic by i p


or topical administration   Daily i p  injection of 0 4 mg 8-MOP to 20 Swiss


albino mice 1 hour before 10-minute exposure to UVA for 6 weeks (total UVA

                     2
exposure 43 joules/cm ) induced skin tumors (mainj.y in ear and eye  regions)  in


virtually all the animals, none of the irradiated control, mice bore such


tumors (60)   Prolonged UVA exposure was less effective because of  severe


phototoxicity.  Significant increase in the incidence of skin tumors was also


observed in mice of several other strains given i p  injection of 0 4 mg 8-MOP


and irradiation with near or intermediate UV (see Table LXXII)   Although no


direct comparative studies have been conducted, topical, application of 8-MOP


appears to be equally effective as (if not more than) i p  injection for


inducing photocarcinogenesis   For example, Nagayo et al  (72) induced skin


tumors in 20 of 20 albino SKH-hairless-1 mice after 40 weeks of daily skin


painting with 40 «g 8-MOP, 30-60 minutes prior to UVA irradiation   For com-


parison, none of the 36 control mice that received 8-MOP alone or UVA exposure


alone developed tumors   Dubertret et ai  (67) found skin tumors in 92% of 40


XVIInc/Z mice given 115 topical applications of 15 ug/un2 8-MOP on  each ear


and irradiated with UV (365 nm)   About 15% of these mice also had  secondary


tumors   Significant increase in the incidence of skin tumors (43%  treated vs


15% control) was observed in Swiss albino mice skin-painted  with as low as 5


ug 8-MOP 15-60 minutes before UV (300-400 nm) exposure (77,  78).

-------
     The importance of the nature, dose and timing of irradiation  in the acti-




vation of 8-MOP to its photocarcinogenic form was first demonstrated by




Griffin and associates.  In contrast to the induction of skin tumors in nearly




100% of mice by a combination of i p  injections of 8-MOP and 10-minute expo-




sures to UVA, a combination of 8-MOP and short wavelength UV (254  run) induced




fewer tumors than exposure to UV-254 nm alone (60)   Prolonging the time of




UVA exposure or reducing the distance between the light source and skin




decreased the tumor incidence because of excessive phototoxicity causing early




death and severe erythemal responses (60)   Delaying the onset of  UVA exposure




from 1 hour to 20-22 hours after 8-MOP injection completely abolished the




photocarcinogenic effect of the treatment (13).  The spectral dependence of




8-MOP photocarcinogenicity has been recently further explored   Grube et al




(73) topically applied 250 ug 8-MOP onto the skin of SKH-hairless-1 mice and




exposed them to UV of three different wavelengths (300-400 nm, 320-400 nm, 365




nm)   The respective skin tumor (squamous or basal cell carcinoma) incidences




were 17/20, 15/19 and 8/19 and the latent period was considerably  shorter  in




the first two groups indicating that UV of broader spectrum is more effective




in eliciting a photocarcinogenic response   Additional exposure to filtered




fluorescent light (280-400 nm) either before or after combined treatment of




8-MOP and UV-365 nm increased the skin tumor incidence to 100% and shortened
                                                                       •«



the latent period, suggesting a possible interaction involving the 365 nm-




induced 8-MOP photoadducts and other molecular Lesions induced by  emissions




between 280-400 nm (see Section 532533).






     In addition to the demonstration of spectral dependence of 8-MOP photo-




care inogenic ity, Grube, Fry and Ley (73, 74) also noted strain difference  in




carcinogenic response to 8-MOP   In marked contrast to the high incidence  of




skin tumors in SHK-hairless-1 mice, 8-MOP plus UV (300-400 nm) produced no
                                      415

-------
skin carcinomas and 8-MOP plus UVA induced only a low  incidence  (4/19) of skin



carcinomas (and after a prolonged latent period)  in HRS/J/Anl  hairless mice



The HRS/J/Anl hairless mice were also less susceptible than  SKH-hairless-1



mice to the carcinogenic effect of a combination  treatment of  8-MOP, UV-365 nm



and filtered fluorescent light (280-400 nm)   However, no significant strain



difference was observed in mice receiving 8-MOP plus UV-365  nm  The strain



difference in response to 8-MOP plus UVA is annulled by treating the HRS/J/Anl



mice with 12-0-tetradecanoylphorbol-13-acetate (TPA),  a typical  tumorigenesis



promoter, suggesting that the strain difference is associated  with a differ-



ence in expression of dormant tumor cells rather  than  with tumor initiation




     532532  CARCINOGENICITY OF PSORALEN AND RELATED COMPOUNDS



STRUCTURE-ACTIVITY RELATIONSHIPS




     In Vivo Can mogenicity   The search for a safer  drug for treatment of



psoriasis, the concern over human exposure to naturally occurring furocoumarin



compounds used in suntan lotion or folk herbal medicine have led to extensive



studies on derivatives and structural analogs of  8-MOP  At  least six deriva-



tives of psoralen and its isomer angelic in have been tested  for  photocarcino-



genic activity in mice   The results of these studies  are summarized in Table



LXXIII




     The carcinogenicity of five furocoumarins has been studied  by Zajdela,



Bisagni and associates (67, 76, 80) using inbred  XVIInc/Z albino mice   The



compounds (dissolved in acetone) were topically applied on the ears of each


                                 *?                                      9
animal over an approximately 1 cm  surface at the dose of 10 or  15 ug/cm , and



irradiated 15 minutes later with UVA or UV-365 nm at the radiation dose of


              2
1 68 joules/cm    High incidence of skin tumors were observed  in mice given



the three bifunctional furocoumarins — psoralen, 8-MOP and  5-MOP (76)   The
                                      416

-------
                                                                                                 P  1  of 2
                                          Table LXXI1I
Comparative Photocarcinogenicity of Psoralen and Related Compounds  in Mice After Skin  Painting'
Irrad lat ion
Compound Mouse strain Dosing regimen
Psoralenb XVIInc/z 115 x 10 jig/cm2
8-MOPc XVIInc/z 115 x 10|Jg/cm2
Albino hairless 34 x 0 01% solution
34 x 0 03% solution
SKH-hairless-1 156-195 x 5 ^ig/<-m2
156-195 x 50 jig/cm2
156-195 x 50 jig/cm2
200 x 40 ug
5-MOP XVIInc/z 115 x 10 /ig/un2
Albino hairless 34 x 0 01% solution
34 x 0 03% solution
3-Carbethoxy- XVIInc/z 106 or 196 x 15 ^ig/cm2
psoralen
SKH-hairless-1 156-195 x 50 wg/cm2
Type
UV-365 nm
UV-365 nm
SSRe
SSRe
UVA
UVA
UVA
UVA
UV-365 nm
SSRe
SSRe
UVA
UVA
Dose
(J/cm2)
1 68
1 68
1 5
1.5
1 0
1 0
2 5
0.72
1.68
1.5
1 5
1 68
1.0
, Skin tumor
Int. idence
97%
92%
100%
100%
65%
70%
54%
100%
85%
100%
100%
0%
0%
Latency
av =
av =
T50
T50
T50
T50
T50
T50
av =
T50
T50

40 wk
58 wk
= ~ 22 wkg
= - 18 wk8
= 56 wk
= 60 wk
= 60 wk
= 24 wk
69 wk
= ~ 25 wk8
= ~ 21 wkg
--
Re ference
(76)
(76)
(75)
(75)
(79)
(79)
(79)
(72)
(76)
(75)
(75)
(66)
(79)

-------
                                             Table LXXIII (continued)
                                                                                                               of 2
Irradiat ion
Compound Mouse strain
Angel ic in XVIInc/z
SKH-hairless-1
5-Methyl- SKH-hairless-1
angel ic in
4,5'-Dimethyl- SKH-hairless-1
angel icin
Aflatoxin Bj SKH-hairless-1
Dosing regimen
106 or
156-195
156-195
156-195
156-195
156-195
196 x 15 fig/cm2
O
x 50 ug/cm
r\
x 5 ^ig/cm
0
x 50 ug/cm
x 50 ug/cm
o
x 50 fig/ cm
200 x 40 ug
Type
UVA
UVA
UVA
UVA
UVA
UVA
UVA
Skin tumor
Dose
2 f
(J/cm ) Incidence Latency
1 68
2 5-5
1 0
1 0
2 5
7 5
0 72
"moderately carcinogenic"
0 20% T5Q = >60 wk
78% T5Q = 53 wk
97% T5Q = 33 wk
97% T50 = 40 wk
100% T5Q = 42 wk
0%
Reference
(cited
_in_ 80)
(79)
(79)
(79)
(79)
(79)
(72)
aAll compounds listed in this table are noncarcinogenic without photoactivation.  At the  radiation doses applied
 (expressed in joules/cm ), irradiation alone is also without carcinogenic effects
 Psoralen is not photocarcinogenic after oral administration (61)
cSee Table LXXII for other careinogenicity studies of 8-MOP.
 3-Carbethoxypsoralen is also not photocarcinogenic after i p. injection (67)
eSSR   simulated solar radiation   Wavelength spectrum includes UVAj UVB, visible and 1R
 Latent period for tumor induction   TCQ = time for appearance of tumor  in 50% of the surviving mice
^Estimated from graph published  in the reference

-------
two raonofunctlonal furocoumarins — angelicin and 3-carbethoxypsoralen (3-CP)




— were found to be "moderately carcinogenic" (cited  in 80) and  inactive (67),




respectively   Based on comparison of tumor incidence and  latent period, the




relative photocarcinogenic potency of the five compounds  follows the order




psoralen > 8-MOP > 5-MOP > angelic in > 3-carbethoxypsoralen (inactive)   This




relative order is generally in good agreement with the relative  photomutagenic




potency of these compounds and led to the suggestion  that mono functional




furocoumarins are less carcinogenic than bifunctional furocoumarins, and may




therefore be potentially safer therapeutic agents for treatment of




psoriasis   The finding of photocarcinogenicity of 5-MOP  is also of great




importance because the compound was, at one time, used an  active ingredient,




to promote pigmentation, in many widely used suntan preparations at concentra-




tions of 12 to 50 ug/ml (5)   Based on photomutagenicity data, 5-MOP was




earlier predicted to be photocarcinogenic (5)   Like  8-MOP, the  photocar-




cinogenicity of 5-MOP can be promoted by the classical tumorigenesis promoter




TPA (see Section 532533) suggesting that photosensitized 5-MOP can




initiate a population of "dormant" tumor cells, which are  normally not




expressed but can persist over a long period of time  and  can progress towards




macroscopic tumors if and when promoted





     The photocarcinogenicity of 5-MOP has been confirmed  by Young et al  (75)




using albino hairless mice   The compound was skin-painted on the flanks of




the mice (applied as a 0 01 or 0 03% solution in a mixture of arachis oil and




isopropyl myristate, the volume applied was not specified), and  photoactivated




by exposure to simulated solar radiation   Compared to 8-MOP, 5-MOP induced




skin tumors with a slightly longer latent period, but the  difference was not




considered significant and the two compounds were considered to  be equally




potent photocarcinogens
                                      417

-------
     The conclusion that bifunctional furocoumarins are more carcinogenic than




their monofunctlonal counterparts has recently been challenged by Mullen et




al  (79) using SKH-hairless-1 mice   Compounds were topically applied (0 05




ml) on the backs of the mice in 0 01 or 0 1% ethanol solutions and photoac-




tivated 45 minutes later with UVA   These investigators did confirm the non-




carcinogenicity of 3-carbethoxypsoraien and the weak carcinogenicity of




angelicin   However, two methyl derivatives of angelicin have been found to be




even more carcinogenic than 8-MOP, inducing more skin tumors with a much




shorter latent period under similar conditions (see Table LXXIII)   The rela-




tive photocarcinogenic potency of the five furocouraarins tested by Mullen et




al  (79) follows the order   5-methylangelicin > 4,5'-dimethylangelicin >




8-MOP > angelicin > 3-carbethoxypsoralen (inactive)   The photocarcinogenicity




of 5-methylangelicin appears to be directly dependent on the dose of the com-




pound applied, whereas that of 8-MOP is not   Increasing the irradiation dose




has no effect on the photocarcinogenicity of 4,5'-dimethylangelicin but




decreases the 8-MOP-induced tumor incidence due to higher phototoxicity   Mice




receiving 0 01 and 0 1% solutions of 8-MOP exhibit severe skin irritation,




scabbing, and epidermal necrosis   No sign of phototoxicity was observed in




mice receiving either of the other four mono functional furocoumarins   These




findings have led the investigators to hypothesize that the DNA photoadducts




formed with both the mono functional furocoumarins (monoadducts) and the




bifunctional furocoumarins (crosslinks as well as monoadducts), undergo an




error-prone excision-repair process which is potentially carcinogenic   The




increased lethality associated with the formation of cross-links by bifunc-




tional furocoumarins may actually reduce the carcinogenic potential of the




affected cells by killing the cells before somatic mutation can take place




The lack of careinogenicity of 3-carbethoxypsoralen was attributed to its
                                      418

-------
photolability   According to some unpublished observations (P C  Joshi and




M A  Pathak, cited in 79), 3-carbethoxypsoralen undergoes rapid  transformation




to a less reactive molecule upon irradiation   It should, therefore, not be




considered as a typical mono functional furocoumarin compound





     In addition to the compounds listed in Table LXXIII, 4, 5' ,8-tnmethyl-




psoralen (trioxsalen) was reported to be carcinogenic  in albino  mice (M A




Pathak, D M  Kramer and T B  Fitzpatrick, cited in 79), the details of the




study were not available at the time of this writing   Aflatoxin B^, a difuro-




coumarocyclopentenone compound (see Section 5311) structurally related to




furocoumarins, is not carcinogenic after topical application  to  SKH-hairless-1




mice followed by UVA irradiation (72)   Actually, aflatoxin B^ can be photo-




activated to bind to DNA   However, the nature of the  photobinding of AFBj  to




DNA differs from that of metabolically activated DNA binding  (see Section




5 3 1.1 4) and is apparently unrelated to the steps leading to carcinogenesis





     In Vitro Carcinogenicity   Four furocoumann compounds have been tested




for in vitro carcinogenic activity by the cell transformation assay   Evans




and Morrow (81) showed that a 8-MOP plus UV treatment  of primary mouse, rat




and hamster embryo cultures and several established cell lines (BHl^i/Cl^-},




NIH/3T3 and Balb/c 3T3) caused a high percentage of cell transformation as




judged by altered cell morphology and growth characteristics   Subcutaneous




inoculation of 8-MOP-plus-UV-transforraed hamster embryo fibroblasts into 2-




day-old newborn Syrian hamsters produced palpable tumors after 3 days   The




same transformed cells grew into spindle cell sarcomas 10-14  days after inocu-




lation into the cheek pouch of adult hamsters   The capability of 8-MOP plus




UV to induce in vitro cell transformation has been confirmed  by  Uwaifo et al




(82) using mouse fibroblast CSH/IOT^ cells   With UVA activation, 8-MOP




induced "type II foci" at concentrations of 1 0, 15 and 2 5 ug/ml   No such







             (                         419

-------
transformation by 8-MOP was observed without UVA exposure   Among  the other




three naturally occurring furocoumarins investigated  in  the same  study,




imperatorin (8-isoamylenoxypsoralen) induced "type  III foci"  at a  concentra-




tion of 10 .ug/ml but was inactive at lower concentrations   Marmesin (which  is




highly cytotoxic) and chalepin did not induce any foci when tested at concen-




trations of up to 1 0 ug/ml and 5 0  ug/ml, respectively   8-Methoxypsoralen,




imperatorin, marmesin and chalepin have all been isolated from two Nigerian




medicinal plants and are suspected to contribute to the  high  incidence of skin




cancer in Nigeria and other tropical countries (82)





     532533  MODIFICATION OF CARCINOGENESIS BY FUROCOUMARIN COMPOUNDS





     There is a scarcity of information on the modification of carcinogenesis




by furocoumarins and the interaction of furocoumarins with other carcinogens




in animals   Fry et_ ^1_  (74) reported that SKH-hairless-1 mice are consider-




ably more susceptible than HRS/J/Anl hairless mice  to skin carcinogenesis




induced by 8-MOP plus UVA   However, treatment of HRS/J/Anl mice with 12-0-




tetradecanoylphorbol-13-acetate (TPA), a  typical tumorigenesis promoter,




eliminates this strain difference rendering the two strains equally suscep-




tible   The promoting effect of TPA  on the photocarcinogenicity of 5-MOP was




also observed by Zajdela and Bisagni (76) using XVIInc/Z mice   Treatment witn




TPA 8 months after the termination of 5-MOP plus UVA  treatment raised the




tumor incidence from 85% to 100% and increased the  percentage of animals with




multiple tumors from 25% to 61%   An analysis of the  distribution  of latent




periods indicated that the additional tumors observed in the  promotion experi-




ment arose from accelerating the emergence of tumors  which would have had very




long latent periods without promotion.  These results indicate that both 8-MOP




and 5-MOP can, after photoactivation, initiate "dormant" tumor cells which can




persist over a long period of time and evolve towards macroscopic  tumors if




and when promoted







                                      420

-------
     As mentioned in Section 532531, Grube et al  (73) demonstrated that

additional exposure to simulated solar radiation either before or after 8-MOP

plus UV-365 run (PUVA) treatment substantially potentiated the photocarcino-

genic effects   Mice treated with the simulated solar radiation plus UV-365 run

alone did not develop any tumors   The simulated solar radiation (280-400 nm)

contains the UVB (280-320 nm) portion which is known to be photocarcinogenic

(see 83)   The results suggest a possible interaction (e g , synergism or

promotion) between PUVA and UVB in photocarcinogenesis   In this respect, it

is interesting to note that Roberts et al  (84) found that mice treated with

subcarcinogenic doses of PUVA were tumor-susceptible when challenged with UVB-

induced "regressor" tumors cells *  The PUVA treatment appeared to generate a

suppressor cell response in mice and apparently allowed the growth of "regres-

sor" tumor cells through immunosuppression   Besides animal data, there are

some epidemiologic evidence and a variety of case reports to indicate that

PUVA therapy may potentiate or promote tumor induction in patients who had a

history of prior exposure to substantial amounts of known carcinogens (see^

Section 53255 1).  Apparent suppression of cell-mediated immune response

in human skin by PUVA treatment also has been observed (85)


     53254  Metabolism and Mechanism of Action


     5 3.2.5 4 1  METABOLISM


     The pharmacokinetic properties of psoralens have been exensively dis-

cussed in a recent symposium (17) with emphasis on development of dosing
*Tumor cells induced by UVB in mice are of  two  immunologic  types  —  the  "pro-
 gressor" cells which grow rapidly when transplanted  into  normal  syngenic
 hosts, and the "regressor" cells which are highly  immunogenic  and  are  usually
 rejected unless  transplanted into immunosuppressed syngenic  hosts
                                      421

-------
regimens to achieve optimal therapeutic efficacy   The metabolism of the two




most widely used psoralen drugs, 8-MOP (methoxsalen) and 4,5' ,8-tnmethyl-




psoralen (trioxsalen) , has been studied by various investigators (25, 86-




90)   Seven metabolites have been isolatd from the urine of SKH-hairless-1




mice or of human volunteers given an oral dose of 8-MOP (25, 87)   Three major




metabolites have been tentatively identified as (a) 8-hydroxypsoralen (only in




human urine), (b) methoxyfurocoumar i<_ acid, the hydrolyzed open-chain form of




8-MOP, which can be cyclized to the lactone form of 8-MOP, and (c) a glucuro-




nide which, upon hydrolysis, regenerates 8-MOP.  Both 8-hydroxypsoralen and




the methoxyf urocourraanc acid are not phototoxic.  A fourth major metabolite




may be   either (a) the 4',5'-epoxide or 4',5'-dihydrodiol derivative of




8-MOP, or (b) the 3,4-epoxide or 3,4-dihydrodiol derivative of 8-MOP, or




(c) 6-carboxymethyl-7-hydroxy-8-methoxycoumarin   The identity of the other




three minor metabolites remains to be elucidated   The metabolites identified




in the urine of dogs given an i v  dose of 8-MOP include the methoxyfurocou—




marie acid, and 6-carboxymethyl-7-hydroxy-8-methoxycoumarin and its conjugate




(88)   In vivo studies showed little or no metabolism of 8-MOP by mouse Liver




homogenates or by microsomes (25)   However, rat liver preparations effec-




tively metabolize 8-MOP in v^tro, and the metabolic rate is enhanced by pre-




treatment of the animals with the typical enzyme inducers, phenobarbital or




^ -naphthoflavone (90)   The metabolism of 4,5',8-trimethylpsoralen in human




volunteers and mice mainly involves oxidation of the 5'-methyl group (86,




89).  The major urinary metabolite is 4,8-dimethyl-5'-carboxypsoralen which is




not phototoxic   Using in vitro systems, it has been established that the




5'-methyl group is first hydroxylated and then further oxidized to carboxy




group.  Both the S'-hydroxy and 5'-carboxy derivatives of 4,5',8-trimethyl-




psoralen may be further hydroxylated   With the possible exception of epoxida-
                                      422

-------
tion, all the above metabolic pathways appear to  represent detoxification




For both 8-MOP and 4,5',8-trimethylpsoralen, there is  no evidence  that




metabolism would enhance the mutagenicity of the  compound  in  the Ames test




(see Table LXXI)   Photoactivation appears to be  the major, if  not  the sole,




pathway in the activation of most psoralens to toxic, mutagenic or  carcino-




genic intermediates





     532542  MECHANISM OF ACTION





     Despite extensive studies (rev  in 8, 12, 14, 19, 20, 25), the mechanism




of carcinogenesis and mutagenesis by psoralens is not  well understood    It




appears that, for most psoralens, at least three  types of mechanisms (a)




intercalation complexrng with DNA, (b) photoreac£ion with  DNA,  and  (c) epi-




genetic mechanisms, either alone or in combination, may contribute  to their




carcinogenic or mutagenic action





     It is now generally believed that intercalation complexing with DNA is




probably responsible for the "dark" mutagenicity  of most psoralens, and  repre-




sents an important initial step in the photobinding of these  compounds to DNA




(see discussion below)   The tricyclic structure  of psoralens,  and  to a  lesser




extent of angelicins, provides a favorable molecular size  and shape for  inter-




calation complexing with DNA   Polynuclear hydrocarbons, as well as poly-




nuclear heteroaromatics structurally related to furocoumarins,  are  known to




intercalate into DNA, resulting in increased solubility, DNA  viscosity^and




melting temperature (Tm) and spectral shifts (see Section  5 1 1 6  2 3 in Vol




IIA and Sections 5 1 5 1 and 5 1.5 2 in Vol  IIB) .





     The evidence supporting the existence of furocoumarin-DNA  intercalation




complexes has been reviewed (1, 20, 91)   The usually  low  solubilities of




furocoumarins in water (see Table LXIX) is significantly enhanced  by the addi-
                                      423

-------
tion of DNA   Likewise, the solubility of DNA  in aqueous solution can be




increased by inclusion of furocoumarin compounds such as 5-MOP   Paralleling




this increased solubility, both the viscosity  and melting  temperature (T ) of




DNA solution are elevated   A spectral shift in the absorption maxima and a




decrease in extinction coefficient are observed after furocoumarin-DNA com-




plexing (1)   There is also evidence that complexing causes  slight unwinding




of the DNA helix (91)   The furocoumarin-DNA noncovalent binding constants of




several furocoumanns have been determined (20, 91)   There  appears  to be an




inverse relationship between the binding (association) constants of  furocou-




marins and their water solubilities, suggesting that mainly  hydrophobic  forces




are involved in the complexes (20) just as with the polynuclear hydrocar-




bons   The stability of the furocoumarin-DNA intercalation complexes is  low




and this is consistent with the observation that most furocoumanns  are  either




inactive or weak, direct-acting, frameshift mutagens when  assayed in the dark




(see Section 532522)   Despite the weak intercalation complexing of the




psoralen molecule with DNA, such intercalation plays an important role if the




molecule is coupled with a reactive functional group   For example,  psoralen-




8-glycidol ether is one of the most potent synthetic mutagens known, substitu-




tion of the psoralen nucleus with a phenyl moiety reduces  the activity by more




than 30 times (see Section 5 3.2 522)   Furocoumanns show a much  lo^er




capacity to form noncovalent complexes with RNA, and photobind to RNA less




effectively than to DNA   Moreover, there is some evidence that furocoumarins




do not intercalate randomly but do so preferentially in regions having a high




density of repetitive alternations of purines  and pyrimidines (20)   A




sequence-specific noncovalent complexing has also been shown to occur before




the covalent binding of the potent carcinogen, aflatoxin B^, to DNA  (see




Section 5 3 1.1 4 2)
                                      424

-------
     The mechanism of photobinding of psoralens and angelic ins to DNA has been




extensively investigated (rev  in 1, 8, 9, 19, 20, 25)   Figure 17 summarizes




the proposed mechanism for bifunctional linear furocoumarins   Once the furo-




coumann-DNA complex is established, covalent bonds between the furocoumarin




molecule and DNA are formed upon irradiation   Firstly, a CA-cycloaddition




reaction involving the 5,6-double bond of pyrimidines and the exicted (triplet




state) 3,4- or 4',5'-double bond of the psoralen molecule occurs, forming 3,4-




or 4',5'-monoadducts   Then, the 3,4- or 4',5'-monoadducts may be further




photoactivated to react with a second pyrimidine base to give a diadduct




When a diadduct is formed with the two strands of DNA, a cross-link is




formed   Photoreaction of angelicins with DNA yields mainly 4',5'-monoadduct




(20, 92) because the angular structure of angeiicins forecloses the possi-




bility of photoreaction of the 3,4-double bond with a second pyrimidine base




in the sister DNA strand   The photoreaction of 3-carbethoxypsoralen is also




expected to yield 4',5'-monoadduct because of the presence of bulky substi-




tuent in the 3-position, however, the identity of this photoadduct has not




been clearly defined (20)   The relative ability of psoralens and angelicins




to photoreact with DNA, in terms of total photobinding (mostly monoadducts




plus some crosslinking), ranks as follows   4,5',8-trimethylpsoralen » 4,5'-




dimethylangelicin > psoralen > 8-MOP > 5-MOP > angelicin > 3-carbethoxy-




psoralen (19, 20).  For crosslinking, the ranking is 4,5'8-trimethylpsoralen




» psoralen > 8-MOP > 5-MOP, whereas 3-carbethoxypsoralen and angelicin show




little or no crosslinking activity (19)   For 4,5',8-trimethylpsoralen, 5-MOP




and 8-MOP, the ratio of raonoadduct to crosslink formation is about 9 1 (20)





     The role of monoadduct formation and crosslinking in photocarcmogenesis




and photomutagenesis by furocoumarin compounds is not clearly understood




Crosslinks are expected to be more effective than raonoadducts for inducing
                                      425

-------
                      0
                    Thymme
                             Intercalation
                                                Psora Ien
complexmg
                                                               0
                                                                  ^0
                 0' -^ -0


               Weak complex
         nr T ""tCHs
         ^o-^V^o
              Weak complex
                          H
                                 Diadduct (Crosslink)
     Fig. 17.  Proposed mechanism for photobinding of psoralen to DNA.  The



first step is believed to involve weak, intercalation complex ing between



psoralen and pyrimid me (e.g., thymine) residues in DMA.  Upon UVA-



irradiation, a CA-cycloaddition reaction takes place, forming 3,4- or



4* ,5'-monoadduct.   Further addition of a second molecule of py raid me to



either monoadduct  gives rise to diadduct (crosslink).  [Modified from B R



Scott, M.A. Pathak, and G.R. Mohn   Mutat.,Res. 39. 29 (1976).]

-------
potentially more genotoxic large-scale changes in the genome (93)   However,




crosslinks are also more difficult to repair (94) and, consequently, are more




likely to lead to cell lethality than monoadducts   Photorautagenicity studies




using various microbial test organisms have consistently shown that bifunc-




tional furocoumarins are more mutagenic than mono functional furocoumarins.




Furthermore, the presence of one or more proficient DNA repair pathways




appears to be a requirement for the expression of photomutagenicity of some




furocoumarins in various microbial test organisms (see Section 532522)




These results suggest that at least one probable mechanism of photomutagenesis




is through "fixation" of the furocoumarin-induced DNA lesions by DNA repair




mechanisms   The usually error-prone .SOS repair mechanisms, triggered by furo-




coumarin-DNA photoadducts (both crosslinks and monoadducts), contribute to




photomutagenesis by promoting cell lethality and by introducing errors during




the repair synthesis   Photocarcinogenicity studies of 8-MOP in mice indicate




that DNA crosslinking does not play a role in carcinogenesis (73) and actually




appears to reduce the carcinogenic potential of 8-MOP by causing cell death




(see Section 532531)   Certain mono functional furocoumanns (e g ,




5-methylangelicin, 4,5'-dimethylangelicin) are in fact more photocarcinogenic




than bifunctional furocoumanns (see Section 532532)   The ranking of




photocarcinogenic potency follows the order   4,5'-dimethylangelicin >




psoralen > 8-MOP > 5-MOP > angelicin > 3-carbethoxypsoralen, which is poorly




related to the relative crosslinking activities of the compounds, but corre-




lates well with their relative ability to form monoadducts   The lack of cor-




relation between mammalian photocarcinogenicity and microbial photomuta-




genicity is intriguing, the possibility that mammalian and microbial cells may




differ in their ability to repair furocoumarin-DNA photoadducts should be




investigated
                                      426

-------
     The mechanisraCs) by which photoadducts of  furocoumarins  initiate  the  pro-




cess of carcinogenesis is not known   It has been  shown  that  both  the  struc-




ture and the template functions of DNA are altered by photoreaction with furo-




coumanns (rev  in 1, 9, 10), and such changes  are expected to be  potentially




genotoxic   The contributory role of DNA repair pathways to photocarcino-




genesis is not clear   The repair of photoadducts  is error-prone and it is




possible that it would introduce errors during  repair synthesis and lead to




somatic cell mutation   However, there is evidence that  the excision repair of




both monoadducts (94, 95) and especially of crosslinks (94) of furocoumarins




is relatively inefficient in mammalian cells    Certain portions of the




mammalian genome (e g», a specific region of DNA ui monkey cells called





-------
gens, by inducing antigen-specific T-suppressor cells   Roberts et al  (84)




have shown that transplanted tumor cells, which normally do not grow  in




immune-proficient mice, grow rapidly in PUVA-treated mice   Thus, it  appears




that the immunosuppression brought about by PUVA treatment promotes dormant




tumor cells or allows their growth and expression   Consistent with the promo-




tion of carcinogenesis by PUVA treatment, Gange (100) showed that PUVA treat-




ment of hairless albino mice induces epidermal ornithine decarboxylase, the




enzyme which is believed to play a role in tumor promotion by the typical




promoter, 12-0-tetradecanoylphorbol-13-acetate (101).  Another possible




indirect mechanism through which PUVA treatment may induce carcinogenesis  is




the induction (activation) of, latent tumor virus(es)   Coohill and Moore (102)




reported that treatment of Simian virus 40-transformed hamster kidney cells




with PUVA leads to the induction of the SV40 virus, an oncogenic DNA  tumor




virus (103)   In addition to the specific indirect mechanisms exemplified




above, psoralens can photoconjugate with proteins leading to the inactivation




of cellular enzymes (19)   It is possible that random inhibition of some key




enzymes can lead to disturbance of gene expression and loss of control of




nuclear function.  Granger et al  (104, 105) have shown recently that PUVA




treatment inactivates two of the three enzymic functions of DNA polyraerase of




Escherichia coll, whether this finding has any relevance to mammalian carcino-




genesis remains to be studied   The reactive oxygen species ( C>2, Oo'j'OH,




etc ) produced by the interaction of triplet state psoralens with molecular




oxygen (see Section 532521) are also potentially carcinogenic    They can




cause direct DNA damage or can initiate membrane lipid peroxidation (19) lead-




ing to the generation of other potentially carcinogenic hydroperoxides and




unsaturated aldehydes (see Vol. IIIA, Section 5.2 1.1 4 1.1)   There  is also




some evidence that the singlet oxygen may be responsible for at least a part




of the photomutagenic effect of 8-MOP (21, 106).

-------
     5_ 3 2 5 5  Environmental Significance





     5 3 2 5.5.1  EPIDEMIOLOGIC EVIDENCE





     Psoralen-containing herbal medicine, in combination with  sunlight, have




been used in the treatment of vitiligo in Egypt and  in  India since  1200-2000




B C  (2)   This form of photochemotherapy has gained widespread  interest  in




modern medicine in the last decade, as oral psoralens plus UVA (PUVA)  therapy




has been found to be highly effective in the treatment  of psoriasis, and




mycosis fungoides as well as vitiligo   Since 1974 several thousand patients




have received PUVA treatment in clinical trials in Europe and  the United




States   The safety of such therapy has been a subject  of great  concern as




animal carcinogenicity data indicate increased incidence of skin cancer in




mice given 8-MOP (topically or intraperitoneally) plus  UVA   Several prospec-




tive epidemiologic studies are currently being undertaken to investigate  the




potential excessive risk of skin cancer in patients  undergoing PUVA therapy




Considering the long latent period required for the  induction  of most  human




cancer, all of these studies can only be considered  to  be at the beginning




stage at the time of this writing.  The interim results of these studies  are




summarized in Table LXXIV   It should be noted that  some of the  patients




selected for PUVA therapy had prior exposure to other treatments (eg  , thera-




peutic doses of X-ray, UVB, arsenical compounds, nitrogen mustard)  that are




themselves potentially carcinogenic   These patients should be considered as a




separate subset in the assessment of carcinogenic risk  of PUVA therapy




Moreover, the susceptibility of individual patients  to  the potential cancer




risk of PUVA therapy is expected to be affected by a variety of  host factors,




such as degree of pigmentation, skin type and DNA-repair efficiency
                                      429

-------
                                                                                                                   1 of 2
                                                       Table LXXIV
            Summary of Epidemiologic Studies on Induction of Skin Cancer in Patients Receiving PUVA Therapy3
Authors
Study group
Cumulat ive
dose
Follow-up
Results
Honigsmann et al. (107)
Wilson (108)
Lassus et al. (109)
Roenigh and Caro (110)
318 psoriasis patients
in Innsbruck, Austria
300 psoriasis patients
in Utah, USA
568 psoriasis patients
in Helsinki, Finland
631 psoriasis patients
in various treatment
centers in the USA
660 J/cm2
<500 J/cra2
27 months
Not specified   4 years
2 1 years
(range,
1-3.6 years)
1,000 J/cm2     4 years
None of 246 patients with no prior
exposure to other carcinogens
developed cancer (cancer incidence
of 0 08 expected)   Four of 72
patients with prior exposure to
arsenic developed skin cancer (an
incidence higher than that of the
general population).

Two patients had skin cancer, one
of the patients had a previous
history of skin cancer

No significant increase in cancer
risk in patients with or without
prior exposure to other
careinogens

Ten patients (3 with prior expo-
sure to ionizing radiation)
developed skin tumors   The inci-
dence of skin cancer in patients
with no prior carcinogen exposure
was not higher than that of the
general population

-------
                                                 Table  LXXIV  (continued)
                                                                                                                   2 of 2

K.
Authors
Wolff, unpublished
Study group
4,600 psoriasis patients
Cumulative
dose
Not specified
Follow-up
6 years
Results
Skin tumor incidence

in 3,450
  results (cited in
  ref  111)
Stern et al.  (112),
  Stern (113)
Harrist et al  (114)
in Europe
1,380 psoriasis patients
in 16 treatment centers
in the U S A
1,600 J/cm2     5 years
230 vitiligo patients
in India
Not specified   55 months
Henseler and Christopher    381 psoriasis patients     Not specified   Since 1976
  (115)                     in the Federal Republic
                            of Germany
patients with no prior exposure to
other carcinogens apparently lower
than expected (0 03 vs  0 08)
Incidence in 1,150 patients with
prior exposure to carcinogen 5
times higher than the general
population

No significant increase in cancer
risk in patients who had no prior
carcinogen exposure and who
received fewer than 80 PUVA treat-
ments   Significant increase (9
times higher) in risk of develop-
ing squamous cell carcinomas in
patients with prior carcinogen
exposure

No evidence of increase in skin
cancer risk

No evidence of increase in skin
cancer risk
 With one exception, all the  studies  involved the use of
 psoralen, and 4,5" ,8-trimethylpsoralen  were used as the
 sunlight
                             8-MOP plus UVA   In the study of Harrist £t__al_  (115), 8-MOP,
                             photosensitizer, and the patients were exposed to natural
 'Cumulative PUVA expressed in terms of Joules/cm2   The data shown were those calculated by Stern (113)

-------
     As Che results summarized in Table LXXIV indicate, there appears to be no




evidence, thus far, that PUVA therapy Leads to increased incidence of skin




cancers in patients who had no history of prior exposure to other carcinogenic




treatment   In fact, in some of these studies, the incidence of skin cancer




was apparently slightly lower than those of the control or general popula-




tion   However, considering the short follow-up time in all these studies and




the low cumulative dose in some of these studies, it would be premature to




draw any conclusion at this point   Even if PUVA therapy is carcinogenic, a




rise in the skin cancer rates in these patients would not be expected until




the 1990s





     In contrast to patients with no prior carcinogen exposure, patients who




had received treatments with X-ray, UVB or arsenicals for their skin disease,




before being subjected to PUVA treatment, developed significantly higher inci-




dences of skin cancer than the control or general population   It is not clear




to what extent the increase may be attributed to PUVA therapy.  Reviewing the




data from five different studies, Halprin et al  (111) calculated that the




maximum incidence of skin cancer in these patients was 2% per year   This rate




was not considered to be any higher than that expected from these patients




without PUVA therapy (116)   Stern (113), however, is of the opinion that pro-




longed exposure to high doses of PUVA can lead to substantial increase in the




risk of developing skin cancer, particularly in patients previously treated




with ionizing radiation   In some patients the tumors appeared within 2 years




after the start of PUVA therapy (112)   After 4 years of follow-up, the




patients were nine times more likely to develop squamous cell carcinomas than




the control group (p < 0 001) and two times more likely to develop basal cell




carcinomas   He cautioned that the negative results seen in other studies




could be due to shorter follow-up time and lower cumulative dose.
                                      430

-------
     In addition to the above studies, there are a number of individual case




reports of rapid induction of skin cancer in patients after receiving PUVA




therapy   In all these cases, the patients either had previous carcinogen




exposure or were believed to be "intrinsically susceptible "  Holier and




Howitz (117) reported that a 43-year-old vitiligo patient, who had been pre-




viously treated with arsenic trioxide 8 years before, developed multiple basal




cell carcinomas 2 months after receiving 8-MOP plus sunlight treatment for 91




days   Verdich (118) detected multiple squamous cell carcinomas in mycosis




fungoides patients who received PUVA therapy after exposure to electron beam




or nitrogen mustard   Stern et al  (119) observed an increase in skin cancer




rate in patients, with previous histories of yery high skin exposure to tar




and far-UV, within 2-7 years after the start of PUVA therapy.  Reed and asso-




ciates (97, 98) found that xeroderma pigmentosum patients (deficient in DNA-




repair enzymes) could develop skin tumors within 1 month after starting PUVA




therapy   The extremely short induction time in these cases, together with




animal data, led Bridges and Strauss (120) to hypothesize that PUVA may act as




a "pseudopromotor," allowing the expression of preexistent neoplastic foci




that would have otherwise remained dormant   This hypothesis was supported by




the observation by Strauss et al. (85) that PUVA therapy can reduce or abolish




cell-mediated immunity in human skin   The totality of the above epidemiologic




evidence and case reports clearly indicate that, at least in certain types of




patients, PUVA therapy can present a substantial increase in cancer risk




Thus, it is imperative to conduct a risk-benefit analysis for each individual




patient selected for PUVA therapy





     5 3 2.5 5 2  ENVIRONMENTAL SOURCES, OCCURRENCE AND EXPOSURE





     Furocoumarin compounds have a wide distribution in the plant kingdom




(rev. in 1, 9, 32, 121, 122)   Over twenty-eight naturally occurring deriva-







                                      431

-------
tives of psoralen or angelicin have been isolated from five major botanical




families   (a) Umbeliferae (parsley family) which includes parsley,  parsnip,




celery, Ammi ma jus (bishop's weed), Angelica ar change lie, a, (b) Rutaceae (rue




family) which includes bergamot fruits, lime, gas plants, cloves, common rue,




(c) Leguminosae (legume family) which includes Psoralea corylifolia,




Xanthoxylum flavum, (d) Moraceae (fig family) which  includes Ficus caria




(fig), and (e) Orchidacea (orchid family)   They are present in the  fruits,




seeds, leaves or roots of the plants   The most frequently found furocoumarins




include 8-MOP (xanthotoxin, methoxsalen), 5-MOP (bergapten) , isopimpinellin




(5,8-dimethoxypsoralen),  4,5',8-triraethylpsoralen (trioxsalen), psoralen,




Vmperatorin (8-isoamylenoxypsoralen) and angelicin (1, 121).   The quantitative




data on concentration are relatively scant   Ivie et al  (123) reported that




parsnip (Past inaca sativ^a) root contains an average of 26 ppm  8-MOP,  10 5 ppm




psoralen and 3 2 ppm 5-MOP   These compounds are not destroyed by usual cook-




ing procedures (boiling or microwave)   Carrot (Daucus carota) root  contains




negligible amounts «0 3 ppm)  of the above three compounds   Trace amounts of




psoralen, 5-MOP, 8-MOP and isopimpinellin were detected in several varieties




of healthy celery (Apium graveolens), their total concentration did  not exceed




1 3 ppm (124)   In contrast to healthy celery, celery infected with  fungus




(e g , Sclerotinia sclerotxorum) produces high concentrations  of psoralens,




perhaps as phytoalexins,  in response to disease infection   Levels of




psoralens (8-MOP, 4,5',8-trimethylpsoralen) in such cases can  reach  hundreds




or perhaps thousands of parts per million (9, 33, 125)





     Human exposure to furocoumann compounds may originate from a variety of




sources which include (a) photochemotherapeutic or medicinal uses, (b) appli-




cation of cosmetics, (c)  occupational exposure and (d) ingestion of  psoralen-




containing foodstuffs   The use of psoralen-containing plant extracts  in
                                      432

-------
photochemotherapy dates back to 1200-2000 B C  when boiled extracts of the




fruit of Annni majus and Psoraiea corylifolia was used ia Egypt aad India,




respectively, along with sunlight, to treat vitiligo (2)   The interest in




this photochemotherapeutic method has surged since 1974 when Parnsh et al




(4) discovered that oral 8-MOP plus UVA (PUVA) therapy is highly effective in




the treatment of psoriasis   Therapy with PUVA has subsequently been found




effective in the treatment of a number of other skin diseases which include




vitiligo, mycosis fungoides (a form of malignant lymphoma which arises in the




skin), polymorphous light eruption, palmar/plantar vesicular dermatitis and




eczema (2)   In some cases, PUVA therapy is the only effective treatment




available   Currently, over 7,500 patients in the United States and Europe are




receiving PUVA therapy in clinical trials   In addition to photochemotherapy,




psoralen-containing plants are also used as folk medicine in some tropical




countries   Concoctions of roots of Clausena anista (wild Rutaceae), Ruta




chalepensis, and Afraegle paniculata (Rutaceae) are used in the treatment of




various diseases ranging from hemorrhoids to measles, scarlet fever, headaches




and heart ailment in Guatemala, Nigeria and other tropical countries (126).




These preparations contain 8-MOP, imperatorin, chalepin and marmesin (see 82)





     Human exposure to furocoumarin compounds may occur through the use of




some cosmetics products which contain oil of bergamot , a citrus extract with




5-MOP or bergapten content (6, 7).  Oil of bergamot was and may still be used




in perfumery in Eau de Cologne which is responsible for the so-called "Eau de




Cologne dermatitis" (erythema and blistering of the skin) in some consumers




(6)   5-Methoxypsoralen was or is perhaps still used as an active ingredient




in many widely used suntan lotions, at concentrations of 12 to 50 -ug/ml, to




promote skin pigmentation (5)   With the demonstration of photocarcinogenicity




of 5-MOP in animals (see Section 532532), it is obvious that such




unnecessary exposure should be avoided

-------
     Besides cosmetics, dermal exposure to psoralens may also occur to profes-




sional fruit or vegetable pickers   Celery workers in Massachusetts (USA) were




reported to develop severe dermatitis after handling rotten celery   It was




subsequently established that extracts of celery rot ("pink rot") contained




high amounts of 8-MOP and 4,5' ,8-tnmethylpsoralen (33) which, in combination




with sunlight, produced the characteristic phototoxic effects (see 34)




Occasional dermatitis has also been reported in lime pickers and in workers in




contact with fig leaves, both of which contain psoralens (1, 6)





     The extent of human exposure to psoralens from food sources is not known




and is dependent on the individual's diet   Human foodstuffs known to contain




psoralerts include figs, parsley, parsnip roots and celery (I, 123, 124)   Ivie




et al  (123) estimated that consumption of 0 1 kg of either raw or cooked




parsnip root, which contains about 40 ppm psoralens, could expose an indivi-




dual to 4 to 5 mg of psoralens   Such exposure might be expected to cause




pharmacological effects in some individuals (the prescribed dosage of




4,5',8-trimethylpsoralen in PUVA therapy for vitiligo is 5 mg)   It is impor-




tant to note that individuals consuming food which contain rotten celery or




parsnip could have substantially higher exposure, because the concentration of




psoralens in such products can reach hundreds and perhaps thousands of parts




per million   Persons in the high risk group (e g , xeroderma pigmentosum




patients, persons with history of prior exposure to carcinogen) should mini-




mize or avoid the consumption of such food products, particularly if they have




a high exposure to sunlight.









                        REFERENCES FOR SECTION 5325









   1   Scott, B R , Pathak, M.A., and Mohn, G.R    Mutat  Res. 39, 29 (1976)
                                      434

-------
 2   Fitzpatrick, T B , and Pathak, M A    NCI Monogr   66,  3  (1984)




 3   El Mofty, A M    J  Roy  Egypt  Med  Assoc  _3]_,  651  (1948)




 4   Parrish, J A , Fitzpatrick, T B , Tannenbaum,  L  ,  and  Pathak, M  A




     New Engl  J  Med  291, 1207 (1974)




 5   Ashwood-Smith, M J , Poulton, G A , Barber, M  , and Mildenberger, M




     Nature (London) 285, 407 (1980)




 6   Stanley, W L , and Jurd, L    J  Agric  Food Chem   19,  1106  (1971)




 7   Porcard, 0 R , and Shubiak, P    J  Assoc  Off  Anal   Chem   57,  145




     (1974)




 8   Song, P S ,  and Tapley, K J  Jr    Photochem   Photobiol   29,  1177




     (1979)




 9   Busby, W F , and Wogan, G N    Psoralens   In  "Mycotoxins and N-Nitroso




     Compounds   Environmental Risks" (R C  Shank,  ed ), Vol   II,  CRC  Press,




     Boc-a Raton,  Florida, 1981, p  105




10   Ciraino, G D ,  Camper, H B , Isaac, S T , and Hearst, J E    Ann   Rev




     Biochem  _54_ (1985, in press)




11   Pearlman, D A , Holbrook, S R , Pirkle, D H ,  and  Kim   S -H    Science




     227. 1304 (1985)




12   Bridges, B A    Environ  Mutagen  _5_, 329 (1983).




13   O'Neal, M A ,  and Griffin, A C    Cancer Res   17,  911  (1957)




14.   Parson, B J    Photochem  Photobiol  32, 813 (1980)




15   International  Agency for Research on Cancer    IARC Monogr  24, 101




     (1980)




16   Scientific International Research (ed.)   "Psoralens in  Cosmetics and




     Dermatology,"  Proceedings of the International Symposium, Paris,  April




     1981, Pergamon, New York, 1982, 447 pp
                                    435

-------
17   Pathak, M A ,  and Dunnick, J K  (eds )   "Photob10 logic, Toxicologic,




     and Pharmacologic Aspects of Psoralens," National Cancer Institute




     Monograph No  66, U S  Government Printing Office, Washington, D C ,




     1984, 261 pp




18   Fowlks, W L    J  Invest  DermatoL  32, 249 (1959)




19   Grossweiner, L.I    NCI Monogr  66, 47 (1984)




20   Rodighiero, G , and Dall'Acqua, F    NCI Monogr. 66, 31 (1984)




21   Joshi, P C  , and Pathak, M A.   Biochem  Biophys. Res   Commun^ 112,  638




     (1983)




22   Averbeck, D ,  Moustacchi, E , and Bisagni, E    Bioihim  Biophys  Acta




     518, 464 (1978)




23   Guiotto, A  , Rodighiero, P , Pastorini, G , Manzini, P  , Bordin, F ,




     Baocichetti, F , Carlassare, F , Vedaldi, D , and Dall'Acqua, F    Eur




     J  Med  Chem  16, 489 (1981)




24.  Windholz, M  (ed )   "The Merck Index," 10th edn , Merck and  Co  ,




     Rahway, New Jersey, 1983




25   Pathak, M A    NCI Monogr. 66^, 41 (1984)




26   Hakim, R E  , Freeman, R G., Griffin, A C , and Knox, J M    J




     Pharmacol  Exptl  Therap  131, 394 (1961)




27   Apostolou, A , Williams, R E , and Comereski, C.R    Drug Chem




     Toxicol  _2, 309 (1979).




28   Herold, M , Berbey, B. , Angignard, D , and Le Due, R.   lexicological




     Studies of the Compound 5-Methoxypsoralen   In "Psoralens in  Cosmetics




     and Dermatology" (Scientific International Research, ed ), Pergamon,




     New York, 1982, p  303.




29.  Dunnick, J K , Davis, W.E. Jr., Jorgenson, T.A., Rosen, V J., and




     McConnell, E E    NCI Monogr  66, 91 (1984).
                                    436

-------
30   Kornhauser, A , Warner, W G  , and Giles, A  L   Jr     NCI  Monogr   66,  97




     (1984).




31   Gange, R W , and Parrish, J A    NCI Monogr   66.  117  (1984)




32   Johnson, A E    Photosensitizing Toxins from  Plants  and Their  Biologic




     Effects.  _In_ "Handbook of Natural Toxins,  Vol.  I,  Plant and  Fungal




     Toxins" (R F  Keeler and AT  Tu, eds ), Marcel Dekker, New  York,  1983,




     p  345




33   Scheel, L , Perone, V , Larkin, R , and Kupel,  R     Biochemistry  2,




     1127 (1963)




34.  Perone, V.B.   The Natural  Occurrence and  Uses  of  the Toxic




     Coumarins   _In_ "Microbial Toxins" (S  Kadis,  A. Ciegler and  S  J  Aj i,




     eds ), Vol  8, Academic Press, New York, 1972




35   Pani, B , Babudri, N , Venturini, S , Tamaro, M.,  Bordin,  F  ,  and




     Monti-Bragadin, C    Teratogen  Carcinogen Mutagen  _1_, 407  (1981)




36   Pani, B., Babudri, N , Gastaldi, S., Carlassare,  F ,  and Guiotto,  A.




     Med  Biol  Environ  11, 447 (1983)




37   Wheeler, L A , DeMeo, M , and Lowe, N    Pho toe hem  Pho t ob101   38.  399




     (1983)




38   Fujita, H    Photochem  Photqbipl  39, 835 (1984).




39   Levin, D E , Hollstein, M., Chnstraan, M.F.,  Schwiers,  E A , and  Ames,




     B N    Proc  Nat  Ac ad  Sci  U.S.. A  79, 7445  (1982)




40.  Venturini, S , Tamaro, M , Monti-Bragad in, C  ,  Bordin,  F , Baccichetti,




     F , and Carlassare, F    Chem -Biol  Interact   30,  203  (1980).




41   Muronets, E M , Kovtunenko, L V , and Kameneva, S.V.    Genetika




     (Moscow) 16, 1168 (1980)




42   Grant, E L , von Borstel, R C , and Ashwood-Sraith, M  J     Environ




     Mutagen  1, 55 (1979)
                                    437

-------
43   Monti-Bragad in, C , Taraaro, M  , Ventunni,  S  ,  Pani,  B  ,  Babudri,  N ,




     and Baccichetti, F    Farmaco  Ed   Sci   36, 551  (1981)




44   Quinto, I , Averbeck, D , Moustacthi, E  , Hrisoho,  Z  ,  and  Moron,  J




     Mutat  Res  136. 49 (1984)




45   Pool, B L , and Deut sch-Wenzel, R P    Aerztl   Kosmetol  _9^> 3^9  (1979)




46   Uwaifo, A 0    J  Toxicol  Environ  Health  13,  521  (1984)




47   Kirkland, D J  , Creed, K L , and Mannisto,  P    Mutat   Re.s   116,  73




     (1983)




48   Rao, T K , Allen, B E , and Cox, J  T     Med   Pediatr  Oncol  10,  100




     (1982)




49.  Scott, B R    Mutat  Res^  40,  167 (1976)




50   Bridges, B A , and Mottershead, R P    Mutat.  Res.  44,  305  (1977)




51   Ivie, G W , MacGregor, J T , and Hammock, B D    Mutat   Res  79,  73




     (1980)




52   Clarke, C H ,  and Wade, M J    Mutat  Res   28,  123  (1975).




53   Ashwood-Smith, M J    Mutat  Res. 58, 23  (1978)




54   Ashwood-Smith, M J , Natarajan, A T , and Poulton,  G  A.   J  Natl




     Cancer Inst  69, 189 (1982)




55   Abel, G , and  Schimmer, 0    Mutat  Res   90,  451 (1981)




56   Schimmer, 0    Mutat  Res. 109, 195 (1983)




57   Averbeck, D ,  Papadopoulo, D., and  Quinto,  I     NCI Monogr. 66,  127




     (1984).




58   Colombo, G , Lev is, A C , and  Torlone, V   Prog  Biochem   Pharmacol




     _!_, 392 (1965)




59   Colombo, G    Cell Res  48, 167 (1967)




60.  Griffin, AC , Hakim, R.E., and Knox, J     J   Invest. Dermatol   31, 289




     (1958)
                                    438

-------
61   Pathak, M A ,  Daniels, F , Hopkins, C E  , and Fitzpatrick,  T  B




     Nature (London) 183, 728 (1959)




62   Hakim, R.E ,  Griffin, A C , and Knox, M    Arch   Dermatol   82,  522




     (1960)




63   Langner, A ,  Wolska, H., Marzulli, F.N , Jablonska,  S  ,  Jarzabek-




     Chorzelska, M , Glinski, W , and Pawinska, M.    J  Invest   Dermatol




     69_, 451 (1977)




64   Pathak, M A ,  and Molica, S.J    Clin  Res. 2_6_,  300A (1978)




65   Griffin, A C    J  Invest  Dermatol  32, 367 (1959)




66   Urbach, F.   J  Invest  Dermatol. 32. 373 (1959)




67.  Dubertret, L , Averbeck, D , Zajdela, F  , Bisagni,  E , Moustacchi, E  ,




     Touraine, A ,  and Latarjet, R    Br. J   Dermatol   101 , 379  (1979)




68   Kripke, M L ,  Morison, W L , and Parrish, J.A.    J  Natl  Cancer  Inst




     68, 685 (1982)




69   Forbes, P D ,  and Urbach, F.   Food Cosmet  Toxicol. 13,  339  (1975)




70   Forbes, P D ,  and Urbach, F    Food Cosmet. Toxicol. 13,  343  (1975)




71   Forbes, P D ,  Davies, R.E , and Urbach,  F.   Food  Cosmet  Toxicol  14,




     303 (1976)




72   Nagayo, K , Way, B H , Tran, RM , and Song, P S    Cancer  Lett   18,




     191 (1983)




73   Grube, D D ,  Ley, R D , and Fry, R J M    Photochem. Photobiol   25,  269




     (1977).




74.  Fry, R.J M ,  Ley, R.D , and Grube, D.D    NCI Monogr  50. 39  (1977)




75   Young, A R ,  Magnus, I.A , Davies, A.C , and Smith,  N  P     Br  J




     Dermatol  108. 507 (1983).




76   Zajdela, F.,  and Bisagni, E    Careinogenesis 2,  121 (1981)
                                    439

-------
77   Santamarie, L , Arnaboldi, A  , Daffara,  P  ,  and  Bianchi,  A    Boll




     Chim  Farm  118. 356 (1979)




78   Santamaria, L , BianchL, A.,  Arnaboldi,  A  ,  and  Daffara,  P    Med




     Biol  Environ  _8, 171 (1980)




79   Mullen, M P , Pathak, M A  , West, J D  ,  Harrist,  T  J  ,  and Dall'Aoqua,




     F    NCI Monogr  66, 205 (1984)




80   Szafarz, D , Zajdela, F ,  Bornecque, C  , and  Barat, N    Photochem




     Photobiol  38, 557 (1983).




81   Evans, D L , and Morrow, K J    J  Invest  Dertnatol   72,  85 (1979).




82.  Uwaifo, A 0 , Billings, P  C ,  and Heidelberger,  C    Cancer Res   43,




     1054 (1983)




83.  Urbach, F    Photocar<_inogensis   _In_ "Derrnatotoxicology"  (F N Marzulli




     and H I  Maibach, eds ), 2nd  edn , Hemisphere, New  York,  1983, p  415




84   Roberts, L K , Schmitt, M  , and Daynes,  R  A.   J  Invest   Dermatol   72,




     306 (1979)




85   Strauss, G H , Bridges, B  A ,  Greaves, M , Hall-Smith,  P., Price, M  ,




     and Vella-Briffa, D.   Lancet  2. 556 (1980)




86   Mandula, B B , Pathak, M A ,  and Dudek,  G.    Science  193. 1131 (1976)




87   Ehrsson, H., Eksborg, S ,  and  Inger, W.    Eur J  Drug  Metab




     Pharmacokinet  _3., 125 (1978).




88   Kolis, S.J , Williams, T H ,  Postma, E  J , Sasso, G J , Confalone,




     P.N , and Schwartz, M A.   Drug  Me tab.  Dispos. _7_»  220  (1979)




89   Mandula, B.B., and Pathak, M.A   Biochem  Pharmacol   28, 127 (1979)




90   Sharp, D E , Mays, D C , Rogers, S.L.,  Guiler, R.C  ,  Hecht, S ,  and




     Gerber, N    Proi  West  Pharmacol  Soc . 27,  255  (1984)




91   Isaacs, S.T , Wiesehahn, G ,  and Hallick,  L  M    NCI  Monogr  66, 21




     (1984)
                                    440

-------
 92   DalL'Acqua, F , Vedaldi, D , Caffien, S  , Guiotto,  A ,  Bordin,  F  ,  and




      Rodighiero, G    NCI Monogr  66. 55 (1984)




 93   Cairns, J    Nature (London) 289. 353 (1981)




 94   Kaye, J , Smith, C A , and Hanawalt, P C.   Cancer  Res   40,  696  (1980)




 95   Zolan, M E, Smith, C A., and Hanawalt, P C    NCI Monogr  66,  137




      (1984)




 96   Baden, H P ,  Pamngton, J M , Delhanty, J D.A  , and  Pathak, M A.




      Biochim. Biophys  Acta 262, 247 (1972)




 97   Reed, WD    Acta Derm. Venereol  (Stockholm) _56_, 315  (1976).




 98   Reed, W.B , Sugarman, G.I , and Mathis, R A.    Arch   Dermatol   113_,




      1561 (1977)




 99   Kripke, M L    J  Nat I. Cancer Inst  69,  171 (1982).




100.  Gange, R W    Br  J. Dermatol. 105, 247 (1981)




101   Boutwell, R K    The Role of the Induction of Ornithine  Decarboxylase




      in Tumor Promotion   In "The Origins of Human Cancer"  (H H  Hiatt,  J D




      Watson and J  A  Winsten, eds ), Cold Spring Harbor,  New  York,  1977,  p




      773




102   Coohill, T P  ,  and Moore, S P    Mutat  Res  113, 431  (1983)




103   Defendi, V.,  and Jensen, F    Science _157_, 703  (1967)




104   Granger, M ,  Toulme, F., and Helene, C    Pho toe hem   Pho t ob101   36_,  175




      (1982).




105   Granger, M ,  and Helene, C.   Photochem. Photobiol.  38.  563  (1983)




106   de Mol, N J ,  Beijersbergen van Henegouwen, G.M.J , Mohn,  G.R  ,




      Glickman, B.W , and van Kleef, P M    Mutat  Res. 82,  23 (1981)




107   Honigsraann, H , Wolff, K , Gschnait, F , et al  .  J.  Am   Ac ad




      Dermatol. _3_,  406 (1980).




108   Wilson, R G.    Dermatology 3, 29 (1980).
                                     441

-------
109   Lassus, A ,  Reunala, T ,  and Idanpaa-Heikkila, J    Acta Derm




      Venereol  (Stockholm) 61, 141 (1981)




110.   Roenigk, H H  Jr ,  and Caro, WA    J  Am  Ac ad  Dermatol  4_, 319




      (1981)




111.   Halprin, KM ,  Coraerford, M , and Taylor, J R    NCI Monogr  66, 185




      (1984)




112   Stern, R S , Thibodeau, L A , Kleinerraan, R A , Parrish, J A ,




      Fitzpatrick, T B ,  and 22 participating  investigators   New Engl  J




      Med  300, 809 (1979)




113   Stern, R S    NCI Monogr  66, 211 (1984)




114.   Harrist, T J ,  Pathak, M A , Masher, D B , and Fitzpatrick, T B    NCI




      Monogr  66,  191 (1984).




115   Henseler, T  , and Christophers, E    NCI Monogr  66^, 217 (1984)




116   Halprin, KM    J  Am  Ac ad  Dermatol  _2, 432 (1980).




117.   Moller, R ,  and Howitz, J.   Arch  Dermatol  112. 1613 (1976)




118   Verdich, J    Arch. Dermatol  115, 1338  (1979)




119.   Stern, R S , Zierler, S , and Parrish, J A    Lancet 1, 732 (1980)




120   Bridges, B A ,  and  Strauss, G H    Nature (London) 283. 523 (1980)




121   Pathak, M ,  Daniels, F  Jr , and Fitzpatrick, T    J  Invest  Dermatol




      39, 225 (1962)




122   Ivie, G W    Toxicological Significance of Plant Fuorooumarins   In




      "Effects of  Poisonous Plants on Livestock," (R F  Keeler, K R  Van




      Kampen and L F  James, eds ), Academic Press, New York, 1978, p  475




123   Ivie, G W.,  Holt, D L , and Ivey, M C    Science 213, 909 (1981)




124   Beier, R C , Ivie,  G W ,  Oertli, E H , and Holt, D L    Food Chem




      Toxicol  _2_i> 163 (1983)




125   Johnson, C , and Brannon, D    Phytochemistry 12_, 2961 (1973)
                                     442

-------
126   Dalziel, JM    "The Useful Plants in West Africa,"  Crown  Agents,




      London, 1937









             SOURCE BOOKS AND MAJOR REVIEWS FOR SECTION  5325









  1   Pathak, M A , and Dunnick, J K  (eds )   "Photobiologic, Toxicologic,




      and Pharmacologic Aspects of Psoralens," Nat  Cancer  Inst  Monograph




      No  66, NIH publ  No  84-2692, U S  Department of  Health and  Human




      Services, Washington, D.C , 1984, 261 pp.




  2   Scientific International Research (ed )   "Psoralens  in Cosmetics and




      Dermatology," Proceeding of the International Symposium, Paris,  April




      1981,  Pergamon, New York, 1982, 447 pp.




  3   Scott, B R , Pathak, M A., and Mohn, G R    Mutat.  Res  39,  29-74




      (1976).




  4   Song,  P -S , and Tapley, K J  Jr    Photochem  Photobiol   29_,  1177-1197




      (1979)




  5   Parsons, B J    Photochem  Photobiol  32_, 813-821  (1980)




  6   International Agency for Research on Cancer   IARC Monogr   24,  101-124




      (1980)




  7.   Busby, W F , and Wogan, G N.   Psoralens   In "Mycotoxins  and  N-Nitroso




      Compounds   Environmental Risks" (R.C  Shank, ed ),  Vol. II,  CRC Press,




      Boca Raton, Florida, 1981, pp. 105-119.




  8.   Bridges, B A    Environ  Mutagen. _5_, 329-339 (1983)




  9.   Cimino, G.D , Camper, H B. , Isaac, S T , and Hearst,  J  E     Ann   Rev




      Biochem. 54 (1985, in press).
                                     443

-------