R E P O R T S
`
`Production of Polyunsaturated
`Fatty Acids by Polyketide
`Synthases in Both Prokaryotes
`and Eukaryotes
`James G. Metz,1* Paul Roessler,2† Daniel Facciotti,2†
`Charlene Levering,2 Franziska Dittrich,2 Michael Lassner,2†
`Ray Valentine,2 Kathryn Lardizabal,2 Frederic Domergue,3
`Akiko Yamada,4† Kazunaga Yazawa,4† Vic Knauf,2† John Browse3*
`
`Polyunsaturated fatty acids (PUFAs) are essential membrane components in higher
`eukaryotes and are the precursors of many lipid-derived signaling molecules. Here,
`pathways for PUFA synthesis are described that do not require desaturation and
`elongation of saturated fatty acids. These pathways are catalyzed by polyketide
`synthases (PKSs) that are distinct from previously recognized PKSs in both structure
`and mechanism. Generation of cis double bonds probably involves position-speci(cid:222)c
`isomerases; such enzymes might be useful in the production of new families of
`antibiotics. It is likely that PUFA synthesis in cold marine ecosystems is accom-
`plished in part by these PKS enzymes.
`
`Very-long-chain PUFAs such as DHA and
`eicosapentaenoic acid (EPA, 20:5v3) have
`been reported from several species of marine
`bacteria,
`including Shewanella sp. (12–14).
`
`Analysis of a genomic fragment (cloned as
`plasmid pEPA) from Shewanella sp. strain
`SCRC2738 identified five open reading frames
`(ORFs) that are necessary and sufficient for
`EPA production in Escherichia coli (13). Sev-
`eral of the predicted protein domains are ho-
`mologs of FAS enzymes, and it was suggested
`that PUFA synthesis in Shewanella involved
`the elongation of 16- or 18-carbon fatty acids
`produced by FAS and the insertion of double
`bonds by undefined aerobic desaturases (15).
`We identified at least 11 regions within the five
`ORFs as putative enzyme domains (Fig. 1A).
`When compared with sequences in the nonre-
`dundant database (16), eight of these were
`more strongly related to PKS proteins than to
`FAS proteins (Fig. 1B). However, three regions
`were homologs of bacterial FAS proteins. One
`of these was similar to triclosan-resistant enoyl
`reductase (ER) from Streptococcus pneumoniae
`(17). Two regions were homologs of the E. coli
`FAS protein encoded by fabA, which catalyzes
`the synthesis of trans-2-decenoyl-ACP and the
`reversible isomerization of this product to cis-
`3-decenoyl-ACP (18). Thus, at least some of
`the double bonds in EPA from Shewanella are
`probably introduced by a dehydrase-isomerase
`mechanism catalyzed by the FabA-like do-
`mains in ORF7.
`To exclude the involvement of an oxygen-
`
`PUFAs are critical components of membrane
`lipids in most eukaryotes (1, 2) and are the
`precursors of certain hormones and signaling
`molecules (3, 4). Known pathways of PUFA
`synthesis involve the processing of the saturated
`16:0 (5) or 18:0 products of fatty acid synthase
`(FAS) by elongation and aerobic desaturation
`reactions (6–8). The synthesis of docosahexean-
`oic acid (DHA, 22:6v3) from acetyl–coenzyme
`A (acetyl-CoA) requires approximately 30 dis-
`tinct enzyme activities and nearly 70 reactions,
`including the four repetitive steps of the fatty
`acid synthesis cycle. PKSs carry out some of the
`same reactions as FAS (9, 10) and use the same
`small protein (or domain), acyl carrier protein
`(ACP), as a covalent attachment site for the
`growing carbon chain. However, in these en-
`zyme systems, the complete cycle of reduction,
`dehydration, and reduction seen in FAS is often
`abbreviated, so that a highly derivatized carbon
`chain is produced, typically containing many
`keto and hydroxy groups as well as carbon-
`carbon double bonds in the trans configura-
`tion. The linear products of PKSs are often
`cyclized to form complex biochemicals that
`include antibiotics, aflatoxins, and many other
`secondary products (9–11).
`
`1Omega Tech, 4909 Nautilus Court North, Boulder,
`CO 80301—3242, USA. 2Monsanto, 1920 Fifth Street,
`Davis, CA 95616, USA. 3Institute of Biological Chem-
`istry, Washington State University, Pullman, WA
`99164 — 6340, USA. 4Sagami Chemical Research Cen-
`ter, 4 — 4 —1 Nishi-Ohnuma, Sagamihara, Kanagawa
`229, Japan.
`
`*To whom correspondence should be addressed. E-
`mail:
`jmetz@omegadha.com ( J.G.M.);
`jab@wsu.edu
`( J. B.)
`†For the present addresses of these authors, contact
`J.G.M.
`
`Fig. 1. Genomics analysis of Shewanella genes encoding enzymes of EPA synthesis. (A) Dark gray
`areas indicate proposed enzymatic domains. Hatched areas designate regions whose amino acid
`sequences are highly conserved amongst the EPA- and DHA-synthesizing proteins of Shewanella,
`Moritella marina (GenBank accession number AB025342.1), and Schizochytrium but for which
`potential enzymatic functions are not evident. Light gray indicates regions whose amino acid
`sequences are not conserved among these organisms. (B) Summary of sequence analysis data for
`the 11 domains designated a to k in (A) (30, 31). a.a., amino acid.
`
`290
`
`13 JULY 2001 VOL 293 SCIENCE www.sciencemag.org
`
`BASF Exhibit 2028, Page 1 of 4
`CSIRO v. BASF, PGR2020-00033
`
`

`

`dependent desaturase in EPA synthesis, we cul-
`tured E. coli harboring the pEPA plasmid
`anaerobically. Anaerobically grown cells accu-
`mulated EPA to the same level as aerobic cul-
`tures (Fig. 2A). When pEPA was introduced
`into a fabB2 mutant of E. coli, which is unable
`to synthesize unsaturated fatty acids that are
`required for growth, the resulting cells lost their
`fatty acid auxotrophy. They also accumulated
`higher concentrations of EPA than other pEPA-
`containing strains (Fig. 2A), suggesting that
`EPA competes with endogenously produced
`monounsaturated fatty acids for transfer to
`glycerolipids. Carbon–13 nuclear magnetic res-
`onance (13C-NMR) analysis of purified EPA
`from the cells grown in [13C]acetate confirmed
`its structure and indicated that this fatty acid
`was synthesized from molecules derived from
`acetate (presumably acetyl-CoA and malonyl-
`
`R E P O R T S
`
`(19). A cell-free homogenate from
`CoA)
`pEPA-containing fabB2 cells synthesized both
`EPA and saturated fatty acids from [1-14C]ma-
`lonyl-CoA (Fig. 2B). High-speed centrifugation
`of the homogenate indicated that saturated fatty
`acid synthesis was confined to the supernatant,
`which is consistent with the soluble nature of
`the type II FAS enzymes (20). Synthesis of
`EPA was found only in the 200,000g pellet
`fraction, indicating that EPA synthesis does not
`rely on FAS or soluble intermediates (such as
`16:0ACP) from the cytoplasmic fraction. Be-
`cause the proteins encoded by the Shewanella
`genes are not particularly hydrophobic, restric-
`tion of EPA synthesis to this fraction may re-
`flect a requirement for a membrane-associated
`acceptor molecule. In contrast to the E. coli
`FAS, EPA synthesis is specifically dependent
`on the reduced form of nicotinamide adenine
`
`Fig. 2. Biochemical analysis of EPA synthesis in E. coli. (A) EPA accumulation in whole cells
`(quanti(cid:222)ed by gas chromatography analysis of methyl-ester derivatives) of E. coli containing a
`plasmid vector alone (control) or vector plus the Shewanella EPA genes [nucleotides 7708 to 35559
`of GenBank accession number U73935.1 in the pNEB vector (New England BioLabs, Beverly, MA)].
`Results were qualitatively similar with a variety of E. coli host strains; control cells never produced
`EPA, whereas the expression of the (cid:222)ve EPA genes resulted in EPA accumulation. Cells were grown
`at 22¡C (13) on media supplemented with glucose. EPA production in a fadE2 mutant (de(cid:222)cient in
`fatty acid degradation) was compared when cells were grown under aerobic ( pEPA-aerobic) and
`anaerobic ( pEPA-anaerobic) growth conditions. Anaerobic conditions were achieved in Gas Pak
`Anaerobic Jars (Becton Dickinson Diagnostic Systems, Sparks, MD). The last column ( pEPA-fabB2)
`shows a level obtained when pEPA was expressed in the fabB2 mutant. (B) Synthesis of EPA and
`saturated fatty acids from [1-14C]malonyl-CoA in subcellular fractions from fabB2 cells expressing
`the EPA genes. Control cells contained a vector alone and were maintained by the supplementation
`of growth medium with oleic acid. Cells were disrupted by passage through a French pressure cell
`and centrifuged (at 8000g for 10 min) to yield a cell-free homogenate (CFH). High-speed
`centrifugation (at 200,000g for 1 hour) was used to generate supernatant (H-S super) and
`membrane (H-S pellet) fractions. The H-S pellet was resuspended in buffer equivalent to the CFH
`volume. Aliquots were incubated in 50 mM phosphate buffer ( pH 7.2), containing 20 mM
`acetyl-CoA, 100 mM [1-14C]malonyl-CoA (0.9 GBq/mol), 2 mM dithiothreitol (DT T), 2 mM NADH,
`and 2 mM NADPH. Lipids were extracted and fatty acids were converted to methyl esters before
`separation by TLC (19). The (cid:222)gure shows radioactivity detected (with a radioanalytic scanner) on
`the TLC plate in the region in which saturated (16:0) and EPA methyl esters migrate. (C) Reductant
`requirement for in vitro EPA synthesis was tested with the H-S pellet fraction described in (B).
`Assays were carried out in reaction buffer containing 2 mM NADH and/or NADPH, or neither, as
`indicated. Incorporation of radioactivity into EPA was measured as described in (B).
`
`dinucleotide phosphate (NADPH) and does not
`require NADH (Fig. 2C). These results are
`consistent with the pEPA genes encoding a
`multifunctional PKS that acts independently of
`FAS, elongase, and desaturase activities to syn-
`thesize EPA directly. Genes with homology to
`the Shewanella EPA gene cluster have been
`found in other PUFA-containing marine bacte-
`ria (21, 22), indicating that the PKS pathway
`may be common in these organisms.
`Schizochytrium is a thraustochytrid marine
`protist that accumulates large quantities of tria-
`cylglycerols rich in DHA and docosapentaenoic
`acid (DPA, 22:5v6) (23). In eukaryotes that
`synthesize 20- and 22-carbon PUFAs by an
`elongation-desaturation pathway, the pools of
`18-, 20-, and 22-carbon intermediates are rela-
`tively large, so that in vivo labeling experiments
`with [14C]acetate reveal clear precursor-product
`kinetics for the intermediates (24). In addition,
`radiolabeled intermediates provided exog-
`enously to such organisms are converted to the
`final PUFA products (25). [1-14C]acetate was
`rapidly taken up by Schizochytrium cells and
`incorporated into fatty acids (Fig. 3A). At 1
`min, DHA contained 31% of the label recov-
`ered in fatty acids, and this percentage remained
`constant during the 10 to 15 min of [1-14C]ac-
`etate incorporation and the subsequent 24 hours
`of culture growth. Similarly, DPA represented
`10% of the label throughout the experiment.
`There is no evidence for a precursor-product
`relation between 16- or 18-carbon fatty acids
`and the 22-carbon polyunsaturated fatty acids
`(25). These results are consistent with rapid
`synthesis of DHA from [1-14C]acetate involv-
`ing very small (possibly enzyme-bound) pools
`of intermediates. A cell-free homogenate de-
`rived from Schizochytrium cultures incorporat-
`ed [1-14C]malonyl-CoA into DHA, DPA, and
`saturated fatty acids (Fig. 3C). The same bio-
`synthetic activities were retained in a 100,000g
`supernatant fraction but were not present in the
`membrane pellet (Fig. 3C). Thus, DHA and
`DPA synthesis in Schizochytrium does not in-
`volve membrane-bound desaturases or fatty
`acid elongation enzymes such as those de-
`scribed for other eukaryotes (7, 8). These frac-
`tionation data contrast with the data obtained
`from the Shewanella enzyme (Fig. 2B) and may
`indicate the use of a soluble acceptor molecule,
`such as CoA, by the Schizochytrium enzyme.
`Additional support for a PKS-based path-
`way was provided from the sequencing of 8500
`randomly selected clones from a Schizochy-
`trium cDNA library (26). Within this data set,
`only one moderately expressed gene (0.3% of
`all sequences) was identified as a fatty acid
`desaturase, although a second putative desatu-
`rase was represented by a single clone (0.01%).
`In contrast, sequences that exhibited homology
`to 8 of the 11 domains of the Shewanella PKS
`genes (Fig. 1) were all identified at frequencies
`of 0.2 to 0.5%. Further sequencing of cDNA
`and genomic clones allowed the identification
`
`www.sciencemag.org SCIENCE VOL 293 13 JULY 2001
`
`291
`
`BASF Exhibit 2028, Page 2 of 4
`CSIRO v. BASF, PGR2020-00033
`
`

`

`of three ORFs containing domains with homol-
`ogy to those in Shewanella (Fig. 4). These
`proteins may constitute a PKS that catalyzes
`DHA and DPA synthesis. The homology be-
`tween the prokaryotic Shewanella and eukary-
`otic Schizochytrium genes suggests that
`the
`PUFA PKS has undergone lateral gene transfer.
`The primary structures of the Shewanella
`and Schizochytrium PKSs do not conform to
`any known class of PKS proteins (9, 10).
`Instead, the data suggest that the PKSs syn-
`thesize PUFAs from malonyl-CoA ( perhaps
`using acetyl-CoA as a primer) with an en-
`zyme complex that carries out iterative pro-
`cessing of the fatty acyl chain but also per-
`forms trans-cis isomerization and enoyl re-
`
`R E P O R T S
`
`duction reactions in selected cycles. Al-
`though the exact
`sequence of
`reactions
`involved in PUFA synthesis remains to be
`determined, schemes can be envisioned that
`accommodate many aspects of the data (27).
`The identification of the PUFA PKSs and
`of putative dehydrase-isomerases may pro-
`vide new tools to engineer the production of
`additional polyketide antibiotics. In addition,
`characterization
`of PUFA synthesis
`in
`Shewanella, Schizochytrium, and their rela-
`tives has implications for understanding food
`web dynamics in aquatic ecosystems (28).
`Because these organisms are important pri-
`mary producers of 20- and 22-carbon PUFAs
`in cold-water oceans (12), the PKS pathway
`
`Fig. 3. Biochemical analysis of
`DHA synthesis in Schizochytrium.
`(A) and (B) In vivo labeling of
`[1-14C]ac-
`Schizochytrium cells.
`etate was supplied to a 2-day-
`old culture as a single pulse at
`zero time. Samples of cells were
`then harvested by centrifugation
`and the lipids were extracted. (A)
`[1-14C]acetate uptake by the
`cells was estimated by measur-
`ing the radioactivity of the sam-
`ple before and after centrifuga-
`tion. (B) Fatty acid methyl esters
`derived from the total cell lipids
`were separated by AgNO3-TLC
`(solvent, hexane:diethyl ether:
`acetic acid, 70:30:2 by volume)
`(32). The identity of the fatty
`acid bands was veri(cid:222)ed by gas
`chromatography, and the radio-
`activity in them was measured
`by scintillation counting.
`(C)
`Synthesis of
`fatty acids from
`[1-14C]malonyl-CoA in subcellu-
`lar fractions from Schizochytrium
`cells. Cells were disrupted in 100 mM phosphate buffer ( pH 7.2), containing 2 mM DT T, 2 mM
`EDTA, and 10% glycerol, by vortexing with glass beads. The cell-free homogenate was centrifuged
`at 100,000g for 1 hour. Equivalent aliquots of total homogenate, pellet (H-S pellet), and super-
`natant (H-S super) fractions were incubated in homogenization buffer supplemented with 20 mM
`acetyl-CoA, 100 mM [1-14C]malonyl-CoA (0.9 GBq/mol), 2 mM NADH, and 2 mM NADPH for 60
`min at 25¡C. Assays were extracted and fatty acid methyl esters were prepared and separated as
`described above before detection of radioactivity with an Instantimager (Packard Instruments,
`Meriden, CT ). The migration of fatty acid standards is indicated.
`
`Fig. 4. Comparison of putative PKS enzyme domains in Shewanella and Schizochytrium ORFs
`(GenBank accession numbers: AF378327, AF378328, and AF378329). Shewanella domains were
`compared with the predicted Schizochytrium gene products by means of the LALIGN program (29)
`with matrix (cid:222)le BLOSUM50 and gap penalties of 214/24 (opening/elongation). Boundaries and
`levels of protein sequence identity for regions identi(cid:222)ed by LALIGN are indicated. KS, 3-ketoacyl
`synthase; MAT, malonyl-CoA; ACP, acyl carrier protein; KR, 3-ketoacyl-ACP reductase; AT, acyl
`transferase; CLF, chainlength factor; ER, enoyl reductase; DH, dehydrase.
`
`may be a significant source of PUFAs for fish
`and mammals.
`
`References and Notes
`1. L. Lauritzen, H. S. Hansen, M. H. Jorgensen, K. F.
`Michaelsen, Prog. Lipid Res. 40, 1 (2001).
`2. M. McConn, J. Browse, Plant J. 15, 521 (1998).
`3. A. Heller, T. Koch, J. Schmeck, K. van Ackern, Drugs
`55, 487 (1998).
`4. R. A. Creelman, J. E. Mullet, Annu. Rev. Plant Physiol.
`Plant Mol. Biol. 48, 355 (1997).
`5. The abbreviation X:Y indicates an acyl group contain-
`ing X carbon atoms and Y cis double bonds; double-
`bond positions of PUFAs are indicated relative to the
`methyl carbon of the fatty acid chain (v3 orv6) with
`systematic methylene interruption of the double
`bonds.
`6. H. Sprecher, Curr. Opin. Clin. Nutr. Metab. Care 2,
`135 (1999).
`7. J. M. Parker-Barnes et al., Proc. Natl. Acad. Sci. U.S.A.
`97, 8284 (2000).
`8. J. Shanklin, E. B. Cahoon, Annu. Rev. Plant Physiol.
`Plant Mol. Biol. 49, 611 (1998).
`9. D. A. Hopwood, D. H. Sherman, Annu. Rev. Genet. 24,
`37 (1990).
`10. R. Bentley, J. W. Bennett, Annu. Rev. Microbiol. 53,
`411 (1999).
`11. T. A. Keating, C. T. Walsh, Curr. Opin. Chem. Biol. 3,
`598 (1999).
`12. D. Nichols et al., Curr. Opin. Biotechnol. 10, 240
`(1999).
`13. K. Yazawa, Lipids 31, S-297 (1996).
`14. F. DeLong, A. A. Yayanos, Appl. Environ. Microbiol.
`51, 730 (1986).
`15. K. Watanabe, K. Yazawa, K. Kondo, A. Kawaguchi,
`J. Biochem. 122, 467 (1997).
`16. S. F. Altschul et al., Nucleic Acids Res. 25, 3389
`(1997). Gapped BLAST and PSI-BLAST protein data-
`base
`search software
`is
`available
`at http://
`www.ncbi.nlm.nih.gov/.
`17. R. J. Heath, C. O. Rock, Nature 406, 145 (2000).
`Comparison of ORF8 peptide with the S. pneumoniae
`ER with the LALIGN program (29) (matrix, BLO-
`SUM50; gap opening penalty, 210; gap elongation
`penalty, 21) indicated 49% similarity over a 386 —
`amino acid overlap.
`18. R. J. Heath, C. O. Rock, J. Biol. Chem. 271, 27795
`(1996).
`19. E. coli (FadE2) cells containing the pEPA plasmid
`were grown to an optical density at 600 nm (OD600)
`of ;1 at 21¡C in 250 ml of liquid medium supple-
`mented with 100 mg of either [1-13C]- or [2-13C]so-
`dium acetate. Cells were collected and lipids extract-
`ed with chloroform and methanol (at a ratio of 2 :1,
`v/v). Fatty acid methyl esters were produced by using
`0.4 ml of 9% sulfuric acid in methanol and 0.2 ml of
`toluene at 90¡C for 1 hour. EPA-methyl esters were
`puri(cid:222)ed by thin-layer chromatography (TLC) [silica
`gel G glass plates were migrated three times in
`hexane and diethyl ether (95:5, v/v)], scraped from
`the plates, and eluted with deuterated chloroform.
`13C-NMR spectra of the samples were obtained on a
`General Electric NMR OMEGA 500 mHz instrument,
`and these spectra con(cid:222)rmed the identity of the EPA
`as 20:5v3. The label from [1-13C]acetate was incor-
`porated into odd-numbered carbons, whereas the
`from [2-13C]acetate was incorporated into
`label
`even-numbered carbons.
`20. K. Magnuson, S. Jackowski, C. O. Rock, J. E. Cronan Jr.,
`Microbiol. Rev. 57, 522 (1993).
`21. E. E. Allen, D. Facciotti, D. H. Bartlett, Appl. Environ.
`Microbiol. 65, 1710 (1999).
`22. M. Tanaka et al., Biotechnol. Lett. 21, 939 (1999).
`23. W. R. Barclay, K. M. Meager, J. R. Abril, J. Appl. Phycol.
`6, 123 (1994).
`24. J. L. Gellerman, H. Schlenk, Biochim. Biophys. Acta
`573, 23 (1979).
`25. We also supplied Schizochytrium cultures with 14C-
`labeled 16:0, 18 :1, or 18:3 fatty acids (as the sodium
`salts). Over 90% of the label was incorporated into
`triacylglycerols and phospholipids, and 10 min after
`
`292
`
`13 JULY 2001 VOL 293 SCIENCE www.sciencemag.org
`
`BASF Exhibit 2028, Page 3 of 4
`CSIRO v. BASF, PGR2020-00033
`
`

`

`R E P O R T S
`
`the addition of substrate, less than 5% of the radio-
`activity remained as unesteri(cid:222)ed fatty acid. These
`results indicate that the fatty acids were taken up by
`the cells and used in the cellular reactions of lipid
`synthesis. However, 24 hours after the start of the
`experiments, 96 to 98% of the radioactivity was
`recovered as the original fatty acid supplied and no
`label could be detected in DHA or DPA.
`26. mRNA was isolated from PUFA-producing Schizochy-
`trium (American Type Culture Collection number
`20888) cells grown in a fermentor, and a cDNA
`library was produced with the GIBCO BRL Superscript
`Plasmid Cloning System with vector pSPORT1.
`Clones were randomly chosen and sequenced from
`the 59 end with a universal sequencing primer.
`27. Supplementary material is available on Science On-
`
`line at www.sciencemag.org/cgi/content/full/293/
`5528/290/DC1. Included is a description of one pos-
`sible reaction scheme. A key feature of this and
`similar schemes is the incorporation of the cis double
`bonds during fatty acyl chain formation by the action
`of the FabA dehydrase/isomerase domains.
`28. D. C. Muller-Navarra, M. T. Brett, A. M. Liston, C. R.
`Goldman, Nature 403, 74 (2000).
`29. W. R. Pearson, Methods Mol. Biol. 132, 185 (2000).
`30. C. Bisang et al., Nature 401, 502 (1999).
`31. We indicate the size of ORF2 as ;1.0 kb in contrast
`to the 0.83 kb given in the GenBank accession num-
`ber U73935.1. Expression of ORF2 in E. coli with the
`ATG at position 9016 as the start codon did not
`produce a functional protein.
`In contrast, a clone
`including the T TG codon at position 9157 did result
`
`in production of a functional protein. Several other
`potential bacterial start codons (i.e., T TG or AT T) are
`present between positions 9157 and 9016, and it is
`possible that one of these codons may represent the
`actual start codon in Shewanella.
`32. M. Miquel, J. Browse, J. Biol. Chem. 267, 1502 (1992).
`33. We thank J. Cronan (Univ. of Illinois) for the gift of E.
`coli strains, B. Shen (Univ. of CA Davis) for the help
`with NMR analyses, and J. Kuner (OmegaTech, Boul-
`der, CO) for assistance with Schizochytrium gene
`sequencing. Supported in part by grants from Mon-
`santo and the U.S. Department of Energy (grant
`DE-FG03-99ER20323) to J.B. and by the Agricultural
`Research Center, Washington State University, Pull-
`man, WA.
`
`5 February 2001; accepted 1 June 2001
`
`Endothelial Apoptosis as the
`Primary Lesion Initiating
`Intestinal Radiation
`Damage in Mice
`Franc(cid:252)ois Paris,1 Zvi Fuks,2 Anthony Kang,1 Paola Capodieci,3
`Gloria Juan,3 Desiree Ehleiter,1 Adriana Haimovitz-Friedman,2
`Carlos Cordon-Cardo,3 Richard Kolesnick1*
`
`Gastrointestinal (GI) tract damage by chemotherapy or radiation limits their
`ef(cid:222)cacy in cancer treatment. Radiation has been postulated to target epithelial
`stem cells within the crypts of Lieberku‹hn to initiate the lethal GI syndrome.
`Here, we show in mouse models that microvascular endothelial apoptosis is the
`primary lesion leading to stem cell dysfunction. Radiation-induced crypt dam-
`age, organ failure, and death from the GI syndrome were prevented when
`endothelial apoptosis was inhibited pharmacologically by intravenous basic
`(cid:222)broblast growth factor (bFGF) or genetically by deletion of the acid sphin-
`gomyelinase gene. Endothelial, but not crypt, cells express FGF receptor tran-
`scripts, suggesting that the endothelial lesion occurs before crypt stem cell
`damage in the evolution of the GI syndrome. This study provides a basis for new
`approaches to prevent radiation damage to the bowel.
`
`The GI syndrome is the main toxicity associat-
`ed with abdominal radiotherapy of human tu-
`mors. It consists of diarrhea, dehydration, en-
`terobacterial infection, and in severe cases, sep-
`tic shock and death (1). According to the pre-
`vailing hypothesis,
`it
`is caused by direct
`damage to a group of stem cells within the
`epithelial rings at positions 4 to 5 from the base
`of the crypts of Lieberku¨hn, resulting in their
`death (1–3). This etiological model is based on
`inference from studies of crypt regeneration
`after injury by using an in vivo clonogenic
`survival assay. Stem cell death is the critical
`element in the evolution of this process, be-
`cause a single surviving stem cell appears suf-
`ficient for reconstitution of a crypt-villus unit
`
`1Laboratory of Signal Transduction and 2Department
`of Radiation Oncology and 3Department of Patholo-
`gy, Memorial Sloan-Kettering Cancer Center, 1275
`York Avenue, New York, NY 10021, USA.
`
`*To whom correspondence should be addressed. E-
`mail: r-kolesnick@ski.mskcc.org
`
`(1). This clonogenic stem cell activity is quan-
`tified by counting regenerating crypts in histo-
`logic sections 3.5 days after irradiation (1).
`Above 8 grays (Gy), dose-dependent stem cell
`death leads to diminution of crypt regeneration,
`until the level of regeneration is insufficient to
`rescue the GI mucosa. In such cases, progres-
`sive denudation of the epithelium leads, by day
`6 to 7 after radiation, to death of mice from the
`GI syndrome. Direct evaluation of stem cell
`function in this process is, however, not feasible
`because there are no markers specific for GI
`stem cells.
`Here, we explore the alternative possibil-
`ity that microvascular endothelium within the
`intestinal mucosa is the actual target of radi-
`ation damage, with stem cell dysfunction as a
`consequence. This hypothesis is supported by
`previous observations that (i) survival factors
`for endothelium [vascular endothelial growth
`factor (VEGF), acidic and basic FGFs, and
`interleukin 11 (IL-11)] protect the gut from
`radiation injury (4–6), and (ii) endothelium is
`
`a principal target for radiation injury to lung
`and brain (7, 8).
`In our initial studies, we used a murine
`whole-body irradiation (WBR) model (9). The
`patterns of lethality and tissue damage are
`shown (Fig. 1) for 8- to 12-week-old C57BL/6
`mice exposed to 12 to 15 Gy WBR, which
`exceeds the minimal dose required to kill all
`exposed animals within 30 days (LD100/30).
`When treated with 12 or 13 Gy, 97% of mice
`died 10 to 13 days after irradiation (median 11
`days; Fig. 1A) and displayed bone marrow apla-
`sia and intact intestinal mucosa (Fig. 1B). Au-
`tologous marrow transplantation, performed 16
`hours after 12 Gy WBR, repopulated the mar-
`row (10) and rescued 90% of the mice. In
`contrast, exposure to 15 Gy resulted in more
`rapid death, with 95% of the mice dying be-
`tween 6 and 8 days (mean 6.8 6 0.99 days,
`median 6 days; Fig. 1A). These animals showed
`denudation of the intestinal crypt and villus
`system but had only partially damaged marrow
`(Fig. 1B), and could not be rescued by autolo-
`gous marrow transplantation (Fig. 1A). Actuar-
`ial survival at 14 Gy differed significantly from
`that after 15 Gy (P , 0.001). At this dose, 25%
`of mice succumbed to death from the GI syn-
`drome and 75% died of marrow failure (P 5
`0.007 versus death from GI syndrome at 15 Gy).
`To determine whether microvascular endo-
`thelial apoptosis correlated with development of
`the radiation-induced GI syndrome, we evaluat-
`ed tissue specimens by hematoxylin and eosin
`(H&E) staining, terminal dexoy transferase-me-
`diated deoxyuridine triphosphate nick end label-
`ing (TUNEL), and in situ labeling with annexin
`V (11), at various times after 8 to 15 Gy WBR.
`Previous studies have shown that radiation in-
`duces early p53-dependent (12) and late p53-
`independent (13) apoptotic responses in crypt
`epithelial cells. However, neither response af-
`fects the ability of stem cells to regenerate crypts
`damaged by doses #15 Gy (13) and hence do
`not appear to be involved in the pathogenesis of
`the GI syndrome. Using a published scoring
`system (14), we confirmed these observations
`and found a maximal epithelial apoptotic index
`of 40.0% at crypt position at 4 hours after 15
`Gy, which decreased to 24.1% at position 5 and
`progressively to 4.1% at position 8. The apopto-
`
`www.sciencemag.org SCIENCE VOL 293 13 JULY 2001
`
`293
`
`BASF Exhibit 2028, Page 4 of 4
`CSIRO v. BASF, PGR2020-00033
`
`