REPORTS

1991); G. M. Kaminska and J. Y. Niederkorn, Investig. malization with the Student's t test. Standard errors Systems, Minneapolis, MN) except lysophosphatidic
Ophthalmoi. Vis. Sci. 34, 222 (1993). were converted t o percentages. acid (Sigma). All proteins and antibodies were exten-
5. A. P. Schachat and R. B. Murphy, Eds., Retina (Mosby, 27. Single-letter abbreviations for the amino acid resi- sively dialyzed against PBS before use in biological
Baltimore, MD, ed. 2, 1994). dues are as follows: A, Ala; D, Asp; E, Glu; F, Phe; G, assays.
6. H:J. Xu et al., Cancer Res. 51, 4481 (1991). ~ l y 1,; Ile; L, L ~M, ~~ ; ~p, pro;
t Q, ; R, 8, ser; 30. Human vitreous fluid was withdrawn from three
7. PEDF was purified from WERI-Rb-27R (6) serum-free cadaveric eyes (refrigerated within 1.4 t o 4.5 hours of
T, Thr; V, Val; X, any amino acid; and Y, Tyr.
conditioned media by sequential steps consisting of death) obtained from individuals without ocular dis-
28, P, J, Polverini, N, P, Bouck, F, Rastinejad, Methods
dialysis (molecular mass cutoff, 30 kD) against dis- ease. Fluid was frozen until used. Fresh vitreous fluid
Enzymol. 198, 440 (1991).
tilled water, 60 t o 95% ammonium sulfate precipi- from bovine and eyes'
tation, step elution from lentil lectin Sepharose 48 29. PEDF antipeptide antibody (anti-PEDF) was raised in
31. We thank A. Mountz for VECF measurements; 8.
rabbits against a peptide containing PEDF amino acids
(Pharmacia) with 0.5 M a-methyl-D-mannopyrano- Kerneddy and the Midwest Eye Banks and Trans~lanta-
327 t o 343, conjugated to Keyhole.limpet hernocyanin,
side, and elution from a HiTrap heparin Sepharose tion Center for human eye tissue; M. K. Francis and V.
and affinity-purified on a peptide column, Polyclonal
column (Pharmacia) with increasing NaCl gradient. Cristofalo for anti-EPC-I; M. O'Reilly and J. Folkman for
antisera against bacterial recombinant PEDFIEPC-l (an-
Purification was monitored by an endothelial cell bovine capillary endothelial cells and angiostatin; and C.
migration assay (26), and the yield was 17.5%. Ed- ti-EPC.l) [B, R, DiPaolo, R, Pignolo, V, Cristofalo,
J, J,
Hawkins, R. O'Grady, and Y. Mu for assistance with
lxp. cell R ~ 220, ~ ,178 (Igg5)] and the antiangiogenic
man degradation of proteolytically derived internal retinoblastomas. Supported by the National Eye Insti-
proteinangiostatin jM. S. OjReilly et Cell 79, 315
peptides of the protein yielded t w o unambiguous tute, the Retina Research Foundation, the National Can-
(1994)1 were giFts,Purchasedreagents included neutral-
sequences (TSLEDFYLDEERTVRVPMMXD and IAQL- cer Institute, and the Chicago Baseball Charities.
izing anti-VECF (Cenzyme, Cambridge, MA), pan anti-
PLTGXM) (27). A BLAST protein homology search
revealed that PEDF contains identical sequences. bodies t o TGFP, and all angiogenic inducers (R & D 15 March 1999; accepted 3 June 1999
8. S. P. Beccerra, in Chemisty and Biology of Serpins, F.
C. Church et al., Eds. (Plenum, New York, 1997), pp.
223-237.
9, J. Tombran-Tink, C, J. Chader, L. V, Johnson, Exp. Eye
Res. 53, 411 (1991); F. R. Steele et al., Proc. Nati.
HMG-1 as a Late Mediator of
Acad. Sci. U.S.A. 90, 1526 (1993).
10. T. Tanawaki, 8. P. Becerra, G, J. Chader, J. P. Schwartz,
1. Neurochem. 64, 2509 (1995); Y. Sugita et ai.,
Endotoxin Lethality in Mice
1. Neurosci. Res. 49, 710 (1997).
11. R. J. Pignolo, V, J. Cristofalo, M. 0 . Rotenberg, 1. 5/01, Haichao Wang,'s3* Ona Minghuang Zhang,3
Chem. 268, 8949 (1993); J. Tombran-Tink et ai., Jaideep M. Vi~hnubhakat,~ Michael ~ m b r e l l i n o ,Jiantu
~ , ~ Che,3
1. Neurosci. 15, 4992 (1995).
12. 8. P. Becerra et ai., 1. Biol. Chem. 268, 23148 (1993); Asia Frazier,'a3 Huan Yang,3 Svetlana I ~ a n o v a , ~
S. P. Becerra e t al., ibid. 270, 25992 (1995). ~ R. M a n ~ g u eEugen
Lyudmila ~ o r o v i k o v a ,Kirk ,~ Fai~t,~
13. D. W. Dawson, 0 . V. Volpert, P. Gillis, unpublished
data. Edward Abraham,5 Jan A n d e r s s ~ n Ulf , ~ Anders~on,~
14. Human PEDF cDNA was engineered by polymerase Patricia E. Molina,' Naji N. Abumrad,' Andrew Sama,'
chain reaction t o encode a COOH-terminal hexa-
histidine tag, cloned into pCEP4 (Invitrogen), and Kevin J. Traceyzs3
transfected into human embryonic kidney cells.
Recombinant PEDF was purified from the condi-
tioned media w i t h the Xpress Protein Purification Endotoxin, a constituent of Cram-negative bacteria, stimulates macrophages to
System (Invitrogen). release large quantities of tumor necrosis factor (TNF) and interleukin-I (IL-I),
15. See supplemental figures, available at w . s c i e n c e m a g .
org/feature/datall040070
which can precipitate tissue injury and lethal shock (endotoxemia). Antagonists
16. For preparation of stromal extract, corneas were of TNF and IL-I have shown limited efficacy in clinical trials, possibly because
freed of associated epithelium and as much of the these cytokines are early mediators in pathogenesis. Here a potential late
endothelium as possible, washed extensively in ice-
cold phosphate-buffered saline (PBS, pH 7.4), and
mediator of lethality is identified and characterized in a mouse model. High
minced into small fragments that were incubated for mobility group-I (HMC-1) protein was found t o be released by cultured mac-
24 hours in PBS containing 0.5 m M phenylmethane- rophages more than 8 hours after stimulation with endotoxin, TNF, or IL-I. Mice
sulfonyl fluoride. The extract was filter sterilized,
stored at -80°C, and tested in migration assays at a
showed increased serum levels of HMC-1 from 8 to 32 hours after endotoxin
final concentration of 10 p g of protein per milliliter. exposure. Delayed administration of antibodies to HMC-1 attenuated endo-
17. Y:Q. W u and 8. P. Becerra, lnvestig. Ophthalmol. Vis. toxin lethality in mice, and administration of HMC-1 itself was lethal. Septic
Sci. 37, 1984 (1996). patients who succumbed to infection had increased serum HMC-1 levels, sug-
18. L. P. Aiello et ai., N. Engi. 1. Med. 331, 1480 (1994);
A. P. Adamis et ai., Am. 1. Ophthalmol. 118, 445 gesting that this protein warrants investigation as a therapeutic target.
(1994).
19. N. Ogata et ai., Curr. Eye Res. 16, 9 (1997); K. Mortality rates for systemic bacterial infec- dotoxin (lipopolysaccharide, LPS) stimulates
Hayasaka et a/., Life Sci. 63, 1089 (1998); S. A.
Vinores et al., 1. Neuroimmunoi. 89, 43 (1998).
tion have not declined significantly, despite the acute, early release of cytoltines such as
20. L. P. Aiello et ai., Proc. Natl. Acad. Sci. U.S.A. 92, advances in antibiotic therapy and intensive TNF and IL-1 P from macrophages, and it is
10457 (1995); A. P. Adamis et ai., Arch. Ophthalmol. care. Bacteria do not directly cause lethal these host products that mediate damage (1).
114, 66 (1996); J. M. Provis et ai., Exp. Eye Res. 65, shock and tissue injury. Rather, bacterial en- Macrophages from C3HIHeJ mice do not re-
555 (1997).
21. L. E. H. Smith et al., Invest. Ophthalmol. Vis. Sci. 35, lease TNF and IL-1when stimulated by LPS;
101 (1994). these animals are resistant to LPS ldthality
22. E. A. Pierce, E. D. Foley, L. E. Smith,Arch. Ophthaimoi. 'Department o f Emergency Medicine and >Depart-
m e n t o f Sureerv.
" < N o r t h Shore Universitv H o s ~ i t a l -
(2). Nonnal, LPS-responsive mice can be
114, 1219 (1996).
23. S. E. Connolly et al., Microvasc. Res. 36, 275 (1988). N e w York University School o f ~ e d i c i n e , ' ~ a n h e s s e t , protected from lethal endotoxemia by thera-
24. M. A. Goldberg, 8. P. Dunning, H. F. Bunn, Science N Y 11030, USA. 3The Picower Institute for Medical peutic agents that selectively inhibit cytokine
242, 1412 (1988). Research, Manhasset, N Y 11030. USA. 4Department action or prevent cytolcine release (3).
25. C. J. Gulledge and M. W. Dewhirst, AnticancerRes. 16, o f Surgery, Klinicum Crosshadern, Ludwig-Maximil-
741 (1996). ians University, Munich, Germany, SDivision of Pul. Translating these pathogenic insights into
26. Migration assays were performed in quadruplicate for monary Sciences and Critical Care Medicine, Univer- clinical therapy has proved difficult, in Part
each sample with bovine adrenal capillary endothelial sity o f Colorado Health Sciences Center, Denver, C O because these "early" mediators (Th-F and
cells or human dermal microvascular endothelial cells
(Clonetics, San Diego, CA) as described (28). To com-
80262, USA. 6Department o f Infectious Disease, Karo-
linska Institute, Huddinge University Hospital, Stock-
I L - ~ )are released within minutes after LPS
'
bine multiple experiments, we first subtracted back- holm, Sweden. 7Department o f Rheumatology, Astrid. ( 4 ) ' Thus, even '
ground migration (Bkgd) toward vehicle (0.1% bovine Linderen's
" Children's H o s ~ i t a l .Karolinska Institute. treatment directed against TNF Or IL-l is
serum albumin) and then normalized data by setting
maximum migration toward inducer alone t o 100%.
Stockholm, Sweden. ineffective (3, 5). Paradoxically, LPS-respon-
All experiments were repeated t w o t o five times. *To w h o m correspondence should be addressed. E- sive mice treated with lethal doses of LPS
Statistics were performed on raw data before nor- mail: hwang@picower.edu succumb at latencies of up to 5 days, long

248 9 JULY 1999 VOL 285 SCIENCE www.sciencemag.org

 REPORTS

after serum TNF and IL-1 have returned to cells, rat adrenal (PC-12) cells, and rat primary mice with a single dose of antibodies to HMG-1
basal levels. Moreover, mice deficient in kidney cells did not release HMG-1 after stim- (anti-HMG-1) 30 min before a lethal dose
TNF die within several days of LPS admin- ulation with LPS, TNF, or IL-1p. Like other (LD,,) of LPS did not prevent LPS-induced
istration (6), suggesting that mediators other macrophage products (for example, TNF, IL- death (Fig. 2A). Based on the kinetics of
than TNF might contribute causally to endo- lp, and macrophage migration inhibitory fac- HMG-1 accumulation in serum (Fig. lC), and
toxin-induced death. tor), HMG-1 lacks a classical secretion signal the relatively short biological half-life of anti-
To identify potential "late" mediators of sequence, so the mechanism of release remains bodies to cytokines (3, 13), we reasoned that
endotoxemia, we stimulated murine macro- to be determined. complete neutralization of a late-appearing me-
phage-like RAW 264.7 cells with LPS and To determine if HMG-1 was released sys- diator might require repeated dosing. Adminis-
analyzed the conditioned culture medium temically during endotoxemia in mice, we tration of anti-HMG-1 in two doses (one 30
by SDS-polyacrylamide gel electrophore- measured serum HMG-1 levels after LPS min before LPS and one 12 hours after LPS)
sis (PAGE). LPS stimulation for 18 hours administration. Serum HMG-1 was readily increased the survival rate of the mice to 30%.
induced the appearance of a 30-kD protein detectable 8 hours after administration of a With three doses of antiserum (-30 min, + 12
that was not apparent at earlier time points. median lethal dose (LD,,) of LPS and was hours, +36 hours), 70% of the treated mice
The NH2-terminal sequence of this late-ap- maintained at peak, plateau levels from 16 to survived, as compared with 0% survival in
pearing factor (Gly-Lys-Gly-Asp-Pro-Lys- 32 hours after LPS treatment (Fig. 1C). controls treated with three matched doses of
Lys-Pro-Arg-Gly-Lys-Met-Ser-Ser) was About 20 to 50 kg of HMG-1 was released preimmune serum (P < 0.05). No late death
identical to murine HMG-1, a 30-kD mem- into the murine circulation within 24 hours occurred over 2 weeks, indicating that anti-
ber of the high mobility group (HMG) after endotoxin administration [assuming a HMG-1 did not merely delay the onset of LPS
nonhistone chromosomal protein family (7, distribution half-life (t,,,) of 3 min and an lethality, but provided lasting protection.
8). Based on the HMG-1 sequence in Gen- elimination t,,, of 20 min]; this is compara- To investigate whether antibody treatment
Bank (accession no. M64986), we designed ble to the quantity of TNF and IL-1 released could be delayed until after administration of
primers and isolated HMG-1 cDNA after by LPS treatment. The kinetics of HMG-1 LPS, we injected anti-HMG-1 beginning 2
polymerase chain reaction (PCR) amplifi- appearance in the blood of LPS-treated mice hours after LPS (followed by additional doses
cation. Recombinant HMG-1 (rHMG-1) differs from that of previously described le- at 12 and 36 hours after LPS). This delayed
protein was expressed in Escherichia coli, thal LPS-induced mediators. treatment conferred significant protection
purified to homogeneity, and used to gen- Passive immunization of unanesthetized against an LD,, of LPS (Fig. 2B). Preimmune
erate polyclonal antibodies (9).
Immunoblot analysis revealed that large
amounts of HMG-1 were released from RAW Fig. 1. (A) Release of HMC-1 from cultured A TNF - +
264.7 cells in a time-dependent manner (Fig. macrophages after stimulation with LPS. Mu-
rine macrophage-like RAW 264.7 cells (Ameri- 30 kD -
lA), beginning 6 to 8 hours after stimulation
can Type Culture Collection, Rockville, Mary-
with LPS. Cell viability, as judged by trypan land) were cultured in RPMl 1640 medium, 12:

:
blue exclusion and lactate dehydrogenase re- 10% FBS, and 1% glutamine. When 70 t o 80% ,U 10-
lease, was unaffected by LPS concentrations confluence was reached, cells were resuspend-
that induced the release of HMG-1, indicat- ed in serum-free OPTI-MEM I medium and '1
ing that HMG-1 release was not due to cell seeded onto tissue culture plates (5 x lo6cells 6:
death. HMG-1 mRNA levels were unaffected
by LPS treatment (Fig. lB), indicating that
per well). After 2 hours, RAW 264.7 cells were
treated with LPS (E. coli 0111:B4, 100 ng/ml)
and proteins in the cell-conditioned medium
.- '-
-= 4-

HMG-1 release is unlikely to be linked to were fractionated by SDS-PACE, excised from

increased transcription of the gene. Stimula- Coomassie- stained SDS-PACE gels, and subject-
tion of RAW 264.7 cells for 18 hours with ed t o NH,-terminal sequencing analysis (Com- Time (hours)
TNF (5 to 100 ng/ml) or IL-1 p (5 to 100
ng/ml) also induced HMG-1 release in a cy-
tokine dose-dependent manner. In contrast,
stimulation with interferon-? (IFN-y) alone
monwealth Biotechnologies, Richmond, Virginia).
Polyclonal antisera against purified recombinant
HMC-1 were generated in rabbits (Biosynthesis,
Lewisville, Texas); immunoblotting showed that
m- Hm-1
antiserum reacted with native HMC-1 released
did not induce HMG-I release, even at con- by RAW cells (inset). HMC-1 levels were mea-
centrations up to 100 Ulml; however, IFN-y sured by optical intensity of bands on immuno-
increased by three- to fivefold the amount of blots with NIH 1.59 image software, with refer-
ence t o standard curves generated with purified 0 8 12 16 (hours)
HMG-1 released by stimulation with either rHMC-1. Data are shown as the mean t- SE (n =
TNF or IL-I (10, 11). Pulse labeling experi- 3). (B) Expression of HMC-1 mRNA in macro-
ments with 35S-methionine revealed that phages. Murine macrophage-like RAW 264.7 cells
most of the HMG-1 released during the first were cultured in RPMl 1640, 10% FBS, and 1%
12 hours after TNF and IFN-y stimulation glutamine, and stimulated with LPS (1 kg/ml) for
was derived from a preformed protein pool. 0,8,12, and 16 hours as indicated. Total RNA was
Radioactivity was incorporated into newly isolated with the SV Total RNA Isolation System
(Promega) and levels of HMC-1 mRNA were de-
synthesized HMG-I from 12 to 36 hours after termined by reverse transcriptase (RT)-PCR with
macrophage stimulation (I 0, 11). the Access RT-PCR System (Promega; P-actin
We next examined the inducible release of primers, 5'-TCATCAACTCTCACCTTCACATCCC-
HMG-1 from other cell types. LPS triggered T-3' and 5'-CCTACAACCATTTCCCCTCCAC-
HMG- 1 release from human primary peripheral CATC-3'; and HMC-1 primers, 5'-ATCCCCA-
AACCACATCCTA-3' and 5'-ATTCATCATCAT- Time (hours)
blood mononuclear cells and primary rnacro- CATCTTCT-3'). (C) Accumulation of HMC-1 in
phages from LPS-sensitive mice (C3WHeN), serum of LPS-treated mice. Male BalbIC mice (20 t o 23 g) were treated with LPS [lo mg/&
but not from macrophages from LPS-resistant intraperitoneally(ip)]. Serum was assayed for HMC-1 by immunoblotting; the detection limit is -50 pg.
C3H/HeJ mice (11, 12). Human primary T Data are shown as the mean +. SE (n = 3).

www.sciencemag.org SCIENCE VOL 285 9 JULY 1999 249

 REPORTS

seruinaeated colltrols all developed lethargy, noglobulin G (IgG) fraction from anti-HMG-1 ble doses of rabbit IgG died (Table 1). Treat-
piloerection, and diarrhea before death, whereas and administered it to mice exposed to an ment with anti-HMG-1 IgG (2 mg per mouse)
anti-HMG-1-treated inice remained well LD,,, of LPS (11, 14). The highest dose of significantly reduced serum HMG-1 levels,
groomed and active, had no diairl~ea,and were anti-HMG-1 IgG tested, 5 mg per mouse, con- whereas no reduction was observed after treat-
viable. To clarify that anti-HMG-1 protected ferred complete protection against an LD,,, of ment with a lower dose of antibodies (0.5 mg
mice from LPS lethality, we purified the Immu- LPS, whereas all control mice given conlpara- per mouse) or with control IgG (5 mg per
mouse). Antiserum against a chemically syn-
" Ab (0 2
a Ab (02
-- 0.4 + 0 4 ml)
0.4 ml)
3 Ab (0 2 mi)
c Preimmune (0.2 + 0 4 + 0 4 ml)
thesized peptide corresponding to the first 12
amino acids of HnlG-1 also significantly atten-
uated the lethality of endotoxemia in mice (15).
A Delayed Ab To determine if HMG-1 was toxic. we
administered highly purified rHMG-1 to un-
5 60- .A
anesthetized BalbIC mice (10 to 50 k g per
m
.? 40 - mouse). Within 2 hours. the mice developed
g .
20-
O-P
o- ? signs of endotoxemia. including lethargy. pi-
loerection, and diarrhea. At higher doses (500
k g per mouse); three of five mice died at 18;
30, and 36 hours after rHMG-1 administra-
04 , 8 ' ,LA tion. Toxicity and lethality were not observed
0 20 40 60 80 100
Time (hours post LPS) in control mice treated with a protein fraction
Time (hours post LPS)
purified from E. coli transformed with a plas-
Fig. 2. (A) Anti-HMC-I protect against o LPS alone v LPS + HMG-1 (1U 0 wg)
A LPS + HFVIG-I (I o tlg) 0 I P S + HFV1G-1 (50 0 pg) mid devoid of HMG-1 cDNA (9), indicating
LPS Lethality in mice. Polyclonal anti-
bodies against rHMC-1 were generated C that the toxicity we observed was specific to
in rabbits, and antiserum was assayed o HMG- 1. To exclude the possibility that en-
for specificity and titer by enzyme- A dotoxin contamination of HMG- 1 prepara-
linked irnrnunosorbent assay and ~ r n - 80 - tions mediated lethality, we injected rHMG-1
munoblotting. Antibodies reacted spe- into LPS-resistant mice. rHMG-1 (500 y g
cifically with HMG-1 and did not cross- 60 - per mouse) was lethal within 16 hours both to
react with LPS, other bacterial proteins,
- 5 C3H!HeJ ( 1 1 = 4) and C3H!HeN ( n = 3)
TNF, or IL-ID. lmmunoblots of lysates
of macrophages or E. coli transformed mice, indicating that HMG-1 itself is toxic
with plasmid containing H M C - I cDNA 2o $ I
v
Y even in the absence of LPS signal transduc-
revealed only one band of immunore- 00- 0 0 tion. When sublethal doses of rHMG-1 were
act~vity.Male Balb/C mice (20 t o 23 g) injected into Balb!C mice together with sub-
were randomly grouped (10 mice per o ' o; ' l o ' $0 ' 80 ' 100 ' 120 lethal doses of LPS, the con~billedchallenge
group) and treated with an LD, of LPS
(25 mg/kg). Anti-HMC-I (Ab) or prelm- Time (hours post LPS) was lethal to 90% of the mice. as compared
mune serum (0.2 m l per mouse, ~ p was
) with 0% lethality in mice exposed to LPS or
administered 30 min before LPS. Additional doses of prelmmune (0.4 ml, ip) or anti-HMG-I (0.4 HMG-1 alone (Fig. 2C). Thus, HMG-1 itself
ml, ip) were administered at 12 and 36 hours after LPS as indicated. (0) Delayed administration of mediates lethality in both LPS-sensitive and
anti-HMC-I protects against LPS lethality in mice. Male Balb/C mice (20 t o 23 g) were randomly LPS-resistant mice.
grouped (seven mice per group) and treated with an LD, of LPS. Anti-HMC-I or preimmune
Animal models of human sepsis. including
serum (0.4 m l per mouse) was administered at 2, 24, and 36 hours after LPS. (C) Administration of
rHMC-1 is lethal t o mice. Recombinant HMC-1 was purified and LPS content determined by the the muline endotoxemia model used here, have
Limulus Amoebocyte Lysate Test (Bio-Whittaker, Walkersville, Maryland). Purified rHMC-1 protein inherent limitatiolls (16). As an initial step in
contained <2.5 ng of LPS per microgram of rHMC-1. Male Balb/C mice (20 t o 23, 10 animals per deteimining whether HMG-1 participates in the
group) were injected with a nonlethal dose of LPS (3.1 mg/kg, ip). Purified rHMG-1 protein was pathogenesis of human sepsis, we studied 8
administered intraperitoneally in the doses indicated at 2, 16, 28, and 40 hours after LPS.
Table 1. Protection against LPS lethality by anti-
Fig. 3. Increased serum HMG-1 160 - HMC-1 IgC. BalbIC mice (male, 20 to 23 g, three
levels in human sepsis. Serum was c to six mice per group) were injected intraperito-
obtained from 8 healthy subjects 140 - n neally with IgC purified from anti-HMG-I or con-
and 25 septic patients infected trol rabbit IgC 30 min before injection of an LD,
:
n

with Cram-positive [Bacillus fragi- $ of LPS. All mice were then treated with additional
lis (one patient), Enterococcus fae- 5 100 - c doses of anti-HMC-I IgC or control IgC at 12 and
calis (one patient), Streptococcus 5 83.7" 24 hours after LPS. Serum HMC-1 levels were
pneumoniae (four patients), Liste- 2 80 - n determined by immunoblots (under denaturing
ria monocytogenes (one patient), conditions) 14 hours after LPS challenge (n = 3
or Staphylococcus aureus (two pa- 60 5 : per group). ND, not determined.
tients)], Cram-negative [E. coli
(seven patients), Klebsiella pneu-
$ 40 -
25.2-
0 I Survival
moniae (one patient), Acineto- 20 - Dose of HMC-1- Serum
2 weeks
bacter calcoaceticus (one patient), m immunoreactive HMG-1
0 7 after LPS
Pseudomonas aeroginosa (one pa- (mg~mouse)
I ~ C (ng/ml)
tient), Fusobacterium nucleatum
(%I
(one patient), Citrobacter freundii Normal Survivors Non-Surviviors 0 1003 2 180 0 (016)
(one patient)], or unidentified Septic patients 0.5 1 0 7 0 2 20 17(116)
pathogens (five patients). Serum 2.0 41 5 2 240 50 (316)*
was fractionated by SDS-PACE, and HMC-1 levels were determined by immunoblotting analysis with 5.0 ND 100 (313)*
reference t o standard curves of purified rHMC-1 diluted in normal human serum; the detection limit is
-50 pg. *P < 0.05 versus normal. **P < 0.05 versus survivors. *P < 0.05 versus treatment with control rabbit IgC.

250 9 JULY 1999 VOL 285 SCIENCE www.sciencernag.org

normal subjects and 25 critically ill septic pa- E. Mayes, E. W. Johns, FEES Lett 122, 264 (1980), L. (1 ( ~ g / m l )was added. The level of HMC-1 In the
t~elltswith bacteremia and sepsis-induced organ Einck and M. Bust~n,Exp. Cell Res. 156, 295 (1985). culture medium was determined 18 hours later by
HMC-1 has also been termed "amphoterin" [J. Park- immunoblotting. HMC-1 was n o t detectable in
dysfunction. HMG-1 was not detectable in the kinen et a/., ]. Biol. Cheni. 268, 19726 (1993)l. culture medium of LPS-stimulated C3H/Hej murine
serum of normal subjects, but significant levels HMC-1 and amphoterin are the same protein; we use rnacrophages; HMC-1 levels reached 1 yg/106 cells
were observed in critically ill patients with sep- the name "HMC-1" t o reflect the original description in the culture medium o f LPS-stimulated C3H/HeN
of this protein as the first member of the HMG murine macrophages.
sis (Fig. 3), and these levels were higher in L. B. Hinshaw et al., Circ. Shock 30, 279 (1990); C.
protein family.
patients who succumbed as compared to pa- HMC-1 was cloned by DNA amplification of the
Nathan and M. Sporn, 1. Cell Biol. 113, 981 (1991);
tients with nonlethal infection. D. C. Remick e t al., Am. 1. Pathol. 136, 49 (1990); R. E.
648-base pair (bp) open reading frame from Rat
Walker e t a/., J. infect. Dis. 174, 63 (1996).
HMG-1 is a highly conserved protein with Brain Quick-Clone cDNA (5 ng: Clontech, Palo Alto,
IgC was purified from anti-HMC-I by Protein C-
CA) with the following primers: 5'-CCCCCGCATC-
>95% amino acid identity between rodent and CTCGACCGAACGATGGCCAAAGGAGATCCTA-3'
Sepharose HiTrap affinity chromatography (Phar-
-
human (1 7-20). It has previously been charac- macia Biotech). Purified IgC fractions were de-
and 5'-CCCGCAACCTTATTCATCATCATCATCTTCT-
salted by ultrafiltration through Centricon-10 (Mil-
terized as a nuclear protein that binds to cmci- 3' (PCR at 94'C for 1 min, 56'C for 2 min, 7Z°C for
lipore), followed by t w o washes w i t h IX phos-
form DNA (21), and as a membrane-associated 45 s; 30 cycles). The 680-bp PCR product was
phate-buffered saline. The specificity o f IgC was
digested with Bam HI and Hind Ill and subcloned into
protein termed "amphoterin" that mediates neu- the Bam HI-Hind Ill cloning sites of the pCAL-n
confirmed by immunoblot analysis of macrophage
lysates, which revealed one band of 30 kD. Anti-
rite outgrowth (19, 20). Extracellular HMG-1 vector (Stratagene, La Jolla, CA). The recombinant
HMC-1 IgC did not cross-react w i t h LPS, TNF, I L - I ,
interacts directly with plasminogen and tissue plasmid was transformed into E. coli BLZI(DE3)pLysS
or bacterial proteins on immunoblots.
(Novagen, Madison, WI), and positive clones were
type plasminogen activator (tPA), which en- confirmed by DNA sequencing of both strands. Trans-
H. Wang and K. 1. Tracey, unpublished observations.
hances plasmin generation at the cell surface; M. P. Fink and 8. 0. Heard, J. Surg. Res. 49, 186
formed cells were induced w i t h isopropyl-o-
(1990); K. J. Tracey, and E. Abraham, SHOCK 11, 224
this system plays a role in extracellular prote- thiogalactopyranoside, and rHMC-1 protein was pu-
(1999); E. A. Deitch, ibid. 9, 1 (1998).
rified with a calmodulin-binding resin column (Strat-
olysis during cell invasion and tissue injury agene). As controls for experiments involving admin-
K. L. Lee e t al., Nucleic. Acids Res. 15, 5051 (1987); S.
(19). In addition, HMG-1 has been suggested to Ferrari, L. Ronfani, 8. Calogero, M. E. Bianchi, I. Biol.
istration of rHMG-1 t o mice, we purified proteins
Chem. 269, 28803 (1994): 8. Ferrari, P. Finelli, M.
bind to the receptor for advanced glycation end from E. coli BLZI(DE3)pLysS that had been trans-
Rocchi, M. E. Bianchi, Genomics 35, 367 (1996).
formed with a plasmid that lacks the HMC-1 cDNA
products (RAGE) (22). insert (pCAL-n). The amount of control material
j. F. Maher and D. Nathans, Proc. Natl. Acad. Sci.
As with other inflammatoly mediators U.S.A. 93, 6716 (1996).
administered t o mice was normalized t o the number
j. Parkkinen etal., J. Biol. Chem. 268, 19726 (1993).
such as TNF and IL-1, there may be protec- of E. coli that produce 0.5 m g of rHMC-I.
M. Salmivirta, H. Rauvala, K. Elenius, M.]alkanen, Exp.
tive advantages of extracellular HMG- 1 when H. Wang and K. J. Tracey, unpublished observations.
Cell Res. 200, 444 (1992).
Supplementary data can be found on Science Online M. E. Bianchi and M. Beltrame,Am. 1.Hum. Genet. 63,
released in nontoxic amounts. Macrophages at www.sciencemag.org/feature/data/l037699.shl. 1573 (1998); D. Landsman and M. Bustin, Bioessays
release HMG-1 when exposed to the early, Macrophages were obtained from the peritoneal cavity 15, 539 (1993); M. Bustin and N. K. Neihart, Cell 16,
acute cytokines, indicating that HMG-1 is of LPS-sensitive (C3HlHeN and Balb/C) or LPS-resistant 181 (1979).
(C3HlHej) mice 4 days after intraperitoneal injection 0 . Hori e t al., 1. Biol. Chem. 270, 25752 (1995); A. M.
also positioned as a mediator of other inflam-
with'2.0 m l of thioglycollate broth (4%; Difco, Detroit, Schmidt e t a/., Diabetes 45, 577 (1996).
matory conditions associated with increased MI). Macrophages were pooled from four mice, resus- We thank C. Dang for technical assistance; j. Eaton,].
levels of TNF and IL-1 (for example, rheu- pended into RPMl 1640, 10% fetal bovine solution Roth, B. Sherry, M. Bukrinsky, and M. Symons for
matoid arthritis and inflammatoly bowel dis- (FBS), and 1% glutamine, and plated at a density of 4 X critical reading of the manuscript; and D. Prieto for
lo6 cells per well in six-well Falcon Primaria tissue administrative assistance.
ease). Indeed, in most inflammatory scenari- culture plates. After 24 hours, the culture medium was
os, LPS is probably not the primary stimulus replaced with serum-free OPTI-MEM-I medium, and LPS 11 December 1998; accepted 3 May 1999
for HMG-1 release; it seems more likely that
TNF and IL-1 function as upstream regula-
tors of HMG-1 release. The delayed kinetics
of HMG- 1 release suggest that serum HMG- 1 Ploidy Regulation of Gene
levels may be a convenient marker of disease
severity. Moreover, the observations that Expression
HMG-1 itself is toxic, and that anti-HMG-1
prevents LPS lethality, point to HMG-1 as a Timothy Galitski,' Alok J. Saldanha,'e2 Cora A. Styles,'
potential target for therapeutic intervention. Eric 5. Lander,',' Gerald R. Fink1,'*

References and Notes

Microarray-based gene expression analysis identified genes showing ploidy-
1. K. J. Tracey e t al., Science 234, 470 (1986); C. A. dependent expression in isogenic Saccharomyces cerevisiae strains that varied
Dinarello, FASEB J. 8, 1314 (1994); D. C. Morrison and in ploidy from haploid t o tetraploid. These genes were induced or repressed in
1. L. Ryan, Annu. Rev. Med. 38, 417 (1987); L. L.
proportion t o the number of chromosome sets, regardless of the mating type.
Moldawer, Crit. Care Med. 22, 53 (1994).
2. B. Beutler, N. Krochin, I. W. Milsark, C. Luedke, A. Ploidy-dependent repression of some G, cyclins can explain the greater cell size
Cerami, Science 232, 977 (1986); 1. W. Killion and associated w i t h higher ploidies, and suggests ploidy-dependent modifications
D. C. Morrison, Infect. Immun. 54, 1 (1986); A. Pol- of cell cycle progression. Moreover, ploidy regulation of the FLO77 gene had
torak e t al., Science 282, 2085 (1998).
3. K. j. Tracey e t a/., Nature 330, 662 (1987): M. Bianchi direct consequences for yeast development.
e t al., 1. Exp. Med. 83, 927 (1996): H. R. Alexander,
C. M. Doherty, C. M. Buresh, D. j. Venzon, j. A. Changes in the number of chromosome sets their ploidy upon fertilization and reduce
Norton, 1. Exp. M e d 173, 1029 (1991); C. Wakaba- occur during the sexual cycle, during meta- their ploidy by half at meiosis. In the devel-
yashi, 1. A. Celfand, 1. F. Burke, R. C. Thompson, C. A.
Dinarello, FASEB 1. 5, 338 (1991). zoan development, and during tumor progres- opment of almost all plants and animals, spe-
4. D. C. Hesse e t al.. Sura. Gvnecol. Obstet. 166. 147 sion. Organisms with a sexual cycle double cialized polyploid and polytene cell types
(1988); H. R. Michie etal.,i. Engl. 1. Med. 318, 1481 arise though endocycles, cell cycles lacking
(1988).
5. M. P. Fink, Crit. Care Med. 23, 989 (1995); E. Abra-
cell division (1). Aberrant cell cycle control
'Whitehead lnstitute for Biomedical Research, 9 Cam-
ham, Chest 113 (suppl.), 2248 (1998). bridge Center. Cambridge, M A 02142, USA. 'Depart-
during tumor progression is thought to result
6. M. W. Marino e t al., Proc. Natl. Acad. Sci. U.S.A. 94, m e n t o f Biology, Massachusetts Institute o f Technol- in polyploidy and altered cell behavior (2).
8093 (1997); F. Amiot, C. Fitting, K. 1. Tracey, J. M.
Cavaillon, F. Dautry, Mol. Med. 3, 864 (1997).
ogy, Cambridge, M A 02139, USA. Cells of different ploidy typically show very
7. M. E. Bianchi, M. Beltrame, C. Paonessa, Science 243, *To w h o m correspondence should be addressed. E- different developmental, morphological, and
1056 (1989);], M. Walker, K. Gooderham, j. R. Hastings, mail: fink@wi.mit.edu physiological characteristics. However, a lack

.sciencemag.org SCIENCE VOL 285 9 JULY 1999 251