The Nature of the Principal Type 1 Interferon-Producing Cells in Human Blood

Abstract

Interferons (IFNs) are the most important cytokines in antiviral immune responses. “Natural IFN-producing cells” (IPCs) in
human blood express CD4 and major histocompatibility complex class II proteins, but have not been isolated and further characterized
because of their rarity, rapid apoptosis, and lack of lineage markers. Purified IPCs are here shown to be the CD4+CD11c− type 2 dendritic cell precursors (pDC2s), which produce 200 to 1000 times more IFN than other blood cells after microbial
challenge. pDC2s are thus an effector cell type of the immune system, critical for antiviral and antitumor immune responses.

Full text

9. Fuel load quantities, divided into standard time-lag
size classes (1 hour, 10 hour, 100 hour, and 1000
hour; corresponding to 0 to 0.6, .0.6 to 2.5, .2.5 to
7.6, and .7.6 cm in diameter), fuel height, and leaf
litter depth were measured along randomly directed
10-m transects at three points within each plot (125,
250, and 375 m). Additional fuel load measurements
were made in the vicinity of all observed fires.
10. At each observation site (n544), roughly 10 m of
fireline was observed every 30 s for several minutes
(average observation time, 7.4 min), and the average
minimum and maximum of each flame characteristic
were recorded. Flareups of short duration or small
area were noted separately.
11. J. K. Agee, Fire Ecology of Pacific Northwest Forests
(Island Press, Washington, DC, 1993).
12. R. C. Rothermel, How to Predict the Spread and
Intensity of Forest and Range Fires (General Technical
Report INT-143, U.S. Forest Service, Ogden, UT,
1983).
13. S. J. Pyne, Introduction to Wildland Fire (Wiley, New
York, 1984).
14. Leaf fall from damaged trees begins within 2 days of
a burn. We have observed contiguous layers of leaf
litter in recently burned forests in close proximity to
smoldering logs. Frequent observations by local res-
idents that forest stands often burn more than once
in a season may be explained by smoldering logs
reigniting fires once litter depths become sufficient
to carry fire through the stand.
15. The mortality of trees (.10 cm dbh) in previously
unburned forests that burned in 1995 was 38% 1
year after the fire and 68% at the end of the second
year. Annual mortality in unburned forest during this
time period was ,1%.
16. D. L. Peterson and K. C. Ryan, Environ. Manag.10, 797
(1986).
17. In forest subjected to fire, vines frequently form a
dense mat at 30 to 200 cm above the ground, and
grasses can form up to 70% of the ground cover.
18. Field studies of forests in 1996 revealed a repre-
sentative plot with only 18 live trees per hectare
(with 302 trees standing dead per hectare). The
fires of 1997 had left only three live trees in this
plot, which was threatened by an oncoming fire,
when recensused.
19. E. Silva, thesis, Pennsylvania State University (1996).
20. M. A. Cochrane and C. M. Souza Jr., Int. J. Remote
Sens.19, 3433 (1998).
21. C. E. Van Wagner, Can. J. For. Res.8, 220 (1978).
22. D. Skole and C. Tucker, Science 260, 1905 (1993).
23. INPE (Instituto Nacional de Pesquisas Espaciais) Des-
florestamento 1995–1997 (Sa˜o Jose´ dos Campos, Sa˜o
Paulo, Brazil, 1997).
24. INPE Desflorestamento, 1993–1994 (Sa˜o Jose´ dos
Campos, Sa˜o Paulo, Brazil, 1996).
25. W. D. Jackson, Proc. Ecol. Soc. Aust.3(1968).
26. M. Mueller-Dombois, in Proceeding of the Confer-
ence—Fire Regimes and Ecosystem Properties (Gen-
eral Technical Report WO-26, U.S. Forest Service,
Honolulu, HI, 1981), pp. 137–176.
27. J. Shukla et al.,Science 247, 1322 (1990).
28. C. E. Van Wagner, Can. J. For. Res.3, 373 (1973).
29. In Paragominas, burns detected in the imagery were
compared with data from landowner questionnaires
(n575) that described fire history from 1982 to
1995. Questionnaire data included 51.4% of the
study region and showed 100% detection of reported
fires that occurred within 1 year of the image date.
Comparisons between the area reported burned by
landowners with data from the imagery classifica-
tions showed that the area burned was systematical-
ly underreported (P,0.001; sign test) by an average
of 43%. Only small fires (,50 ha) were overestimat-
ed by landowners.
30. We thank C. Uhl, A. Taylor, G. P. Patil, F. Williams, and
four anonymous reviewers for comments on earlier
versions of the manuscript. This research was funded
by a grant from the PPG7 “Programa de Pesquisa
Dirigida” (MMA/MCT/FINEP). Additional financial
support was provided by NSF, NASA, and the U.S.
Agency for International Development.
16 February 1999; accepted 16 April 1999
The Nature of the Principal
Type 1 Interferon–Producing
Cells in Human Blood
Frederick P. Siegal,
1*
Norimitsu Kadowaki,
2
Michael Shodell,
3
Patricia A. Fitzgerald-Bocarsly,
4
Kokila Shah,
1
Stephen Ho,
2
Svetlana Antonenko,
2
Yong-Jun Liu
2*
Interferons (IFNs) are the most important cytokines in antiviral immune re-
sponses. “Natural IFN-producing cells” (IPCs) in human blood express CD4 and
major histocompatibility complex class II proteins, but have not been isolated
and further characterized because of their rarity, rapid apoptosis, and lack of
lineage markers. Purified IPCs are here shown to be the CD4
1
CD11c
2
type 2
dendritic cell precursors (pDC2s), which produce 200 to 1000 times more
IFN than other blood cells after microbial challenge. pDC2s are thus an
effector cell type of the immune system, critical for antiviral and antitumor
immune responses.
Interferons were discovered in the 1950s as
factors rapidly produced by virus-infected cells
that enable neighboring cells to resist virus
infection (1). IFN-a(leukocyte IFN) and IFN-b
(fibroblast IFN), the two type 1 antiviral IFNs,
are distinct from type 2 IFN-gproduced by
effector T cells. Specialized leukocytes, the
“natural IFN-producing cells” (IPCs), were
shown to be the chief IFN-aproducers in re-
sponse to enveloped viruses, bacteria, and tu-
mor cells (2–14). IPCs express CD4 and major
histocompatibility complex (MHC) class II, but
lack hematopoietic-lineage markers (2–14).
The nature of IPCs—whether they represent
dendritic cells (6,12,14) or cells of a distinct
lineage (7,9)—has been controversial. There is
a progressive loss of CD4
1
T lymphocytes and
functional IPCs during human immunodefi-
ciency virus (HIV) infection (15,16). Preser-
vation of IPCs is associated with protection
from opportunistic infections, suggesting the
importance of IPCs in the host defense (16).
A plasmacytoid cell type from human ton-
sils and blood that lacks lineage markers also
expresses CD4 and MHC class II (17–21).
These cells differentiate into type 2 dendritic
cells (DC2s) when cultured with interleukin-3
(IL-3) and CD40 ligand (19,21). Unlike mono-
cyte-derived type 1 dendritic cells (DC1s) that
induced type 1 T helper cell (T
H
1) differentia-
tion, DC2s induced type 2 T helper cell (T
H
2)
differentiation (21). Here we investigated
whether DC2 precursors (pDC2s) represent
IPCs. Human peripheral blood cells were sep-
arated into the following populations (19,21):
(i) monocytes (over 90% purity), obtained by
centrifugation through 52% Percoll, then
magnetic bead depletion of B, T, and natural
killer (NK) cells; (ii) CD4
1
CD3
2
CD11c
1
immature DCs (99% purity) and (iii) CD4
1
CD3
2
CD11c
2
pDC2s (99% purity), ob-
tained by magnetic bead depletion of B, T,
NK cells, and monocytes, followed by fluo-
rescence-activated cell sorting (FACS) (Fig.
1, A and B); (iv) pDC2-depleted blood mono-
1
Saint Vincents Hospital and Medical Center, New
York, NY 10011, and Long Island Jewish Medical
Center–Albert Einstein College of Medicine, New
Hyde Park, NY 11040, USA.
2
DNAX Research Insti-
tute, Palo Alto, CA 94304, USA.
3
The CW Post Cam-
pus of Long Island University, Brookville, NY 11548,
USA.
4
University of Medicine and Dentistry, New
Jersey Medical School, Newark, NJ 07103, USA.
*To whom correspondence should be addressed.
Table 1. Precursor DC2 cells are the natural IFN-producing cells. Cells (2 310
5
) were cultured for 24
hours with HSV. Without HSV, IFN activity from different cell types was less than 12.5 U/ml (23). PBMC:
total blood mononuclear cells; pDC2-dep: blood mononuclear cells positively selected for expressing CD3,
CD11c, CD19, CD14, and CD56; pDC2-enrich: blood mononuclear cells that were depleted of cells
expressing CD3, CD19, CD14, and CD56; pDC2: FACS-sorted CD4
1
CD11c
2
lin
2
cells; CD11c
1
DC:
FACS-sorted CD11c
1
lin
2
immature DCs; Mo: monocytes; DC1: monocyte-derived DCs after 6 days of
culture with either granulocyte-macrophage colony-stimulating factor (GM-CSF) 1IL-4 or GM-CSF 1
IL-4 1CD40 ligand (21); DC2: pDC2-derived DCs after 6 days of culture with IL-3 or IL-3 1CD40 ligand
(23). ND, not determined.
IFN (U/ml)
PBMC pDC2-dep pDC2-enrich pDC2 CD11c
1
DC Mo DC1 DC2
Exp. 1 500 ND 2,800 89,800 120 ND ,12.5 1,100
Exp. 2 40 ,12.5 180 20,000 ,12.5 350 ,12.5 ,12.5
Exp. 3 700 40 2,800 638,000 70 90 ND ND
REPORTS
www.sciencemag.org SCIENCE VOL 284 11 JUNE 1999 1835

nuclear cells; and (v) pDC2-enriched blood
mononuclear cells (19). pDC2s have a plas-
macytoid morphology, with rough endoplas-
mic reticulum and Golgi apparatus (Fig. 1, C
and D). The CD11c
1
blood immature DCs
display short dendrites (Fig. 1, E and F). The
frequency of pDC2s in human blood mono-
nuclear cells is less than 0.5% and increased
to 3 to 10% after magnetic bead depletion of
lineage-positive cells. The pDC2-depleted
population contains B, T, NK cells, mono-
cytes, and DCs. These cell populations were
exposed to ultraviolet (UV)-irradiated herpes
simplex virus (HSV) for 24 hours (22), and
IFN in the culture supernatant was measured
by a bioassay (23).
IFN production by total peripheral blood
mononuclear cells (PBMCs) from three donors
was 40, 500, and 700 international units (IU)
per 2 310
5
cells (Table 1). There was a four to
six times increase in IFN production from
pDC2-enriched blood mononuclear cells (180
to 2800 IU per 10
5
cells) and a 180 to 911 times
increase in IFN generation from purified
pDC2s (20,000 to 638,000 IU per 10
5
cells).
pDC2-depleted PBMCs, immature CD11c
1
DCs, monocytes, and monocyte-derived
DC1s produced little or low levels of IFN.
The ability of pDC2s to produce IFN was
decreased after maturation into DC2s by cul-
ture with IL-3 or IL-3 plus CD40L for 6 days
(Table 1). The geometric mean IFN-agener-
ation by pDC2s was ;1 IU per cell, similar to
previous estimates (2). Immunoperoxidase
staining for human IFN-aconfirmed that
most pDC2s contained IFN-aprotein after
6 hours of exposure to HSV (Fig. 2) (22).
Analysis of IFN-aand IFN-bmRNA by
polymerase chain reaction (PCR) showed that
among human blood cells, pDC2s were mak-
ing the most IFN-aand IFN-bmRNA (Fig.
3) (24). Thus, the blood cells responsible for
IFN generation in response to HSV, previ-
ously known as the IPCs, are actually the
DC2 precursors. These cells can be traced
and isolated by their expression of CD4 or
IL-3 receptor after depletion of cells express-
ing lineage markers and CD11c (19,21).
The purified IPCs also produced high levels
of IFN in response to Sendai virus and heat-
killed Staphylococcus aureus, confirming the
previous studies on PBMCs and partially puri-
fied IPCs (2–14). The ability of UV-irradiated
virus and heat-killed bacteria to induce IFN
production by IPCs suggests that viral infection
is not required for triggering IFN production.
The rapid production of IFN by IPCs in the
absence of other cells suggests that IPCs repre-
sent an effector cell type of the innate immune
system. We propose that the IPCs/pDC2s
should be included in the hematopoietic devel-
opmental chart as a distinct cell lineage. They
function as professional IFN-producing cells at
the precursor stages and as professional anti-
Fig. 1. Tracing and isolation of IPCs/pDC2s from human peripheral blood. CD3
1
T cells, CD19
1
B
cells, CD16
1
and CD56
1
NK cells, and CD14
1
monocytes were depleted from blood mononuclear
cells by immunomagnetic beads (Dynabeads M-450; Dynal, Oslo, Norway). The cells were stained
with anti-CD4-Tricolor (Immunotech, Marseille, France), anti-CD11c-PE (Becton Dickinson, San
Jose, California), and a mixture of fluorescein isothiocyanate–labeled antibodies to CD3, CD15,
CD16, CD20, CD57 (Becton Dickinson), CD14 (Coulter, Miami, Florida), and CD34 (Immunotech).
Within the lineage-negative population (A), CD4
1
CD11c
2
IPCs and CD11c
1
immature DCs were
isolated (B). IPCs are plasmacytoid by Giemsa staining (C) and contain rough endoplasmic
reticulum and Golgi apparatus under transmission electron microscopy (D). The CD11c
1
blood
immature DCs display dendrites (Eand F).
A
B
Fig. 2. Immunoperoxidase staining for IFN-a
(21,22). (A) Purified CD4
1
CD11c
2
lin
2
IPCs
were stimulated with HSV for 6 hours. (B) IPCs
cultured in medium for 6 hours. The isotype
controls for the primary antibody in both cell
preparations show no staining.
REPORTS
11 JUNE 1999 VOL 284 SCIENCE www.sciencemag.org1836

gen-presenting type 2 DCs upon terminal
differentiation.
Type 1 IFNs have pleiotropic effects on the
immune system, including up-regulation of
MHC class I on all cell types and activation of
macrophage and NK cells (2). IFNs are also
critical in the activation and survival of both
CD4
1
and CD8
1
T cells (25,26). Now with
the ability to trace and isolate IPCs, it should be
possible to directly study the interaction be-
tween IPCs and other cell types within the
immune system. IFN-ahas been widely used
for treating hepatitis B and C as well as various
cancers. A progressive loss of IPCs has been
observed during HIV infection, suggesting that
IPCs may represent targets for HIV-mediated
infection and deletion. The present study pro-
vides an approach to directly monitor the num-
ber and functional state of IPCs in these pa-
tients. The ability to purify and culture IPCs in
vitro will allow further studies on the molecular
mechanisms that control the survival and
growth of IPCs and their production of IFN,
which may lead to novel therapies for patients
with viral infections and cancer.
References and Notes
1. A. Isaacs and J. Lindemann, Proc. R. Soc. London 147,
258 (1957).
2. P. Fitzgerald-Bocarsly, Pharmacol. Ther. 60,39
(1993).
3. H. Kirchner et al.,Immunobiology 156, 65 (1979).
4. H. H. Peter et al.,Eur. J. Immunol. 10, 547 (1980).
5. J. Abb, H. Abb, F. Deinhart, Clin. Exp. Immunol. 52,
179 (1983).
6. B. Perussia, V. Fanning, G. Trinchieri, Nat. Immun. Cell
Growth Regul. 4, 120 (1985).
7. J. Chehimi et al.,Immunology 68, 488 (1989).
8. K. Sandberg et al., Scand. J. Immunol. 34, 565 (1991).
9. S. E. Starr et al.,Adv. Exp. Med. Biol. 329, 173 (1993).
10. H. Svensson et al., Scand. J. Immunol. 44, 164 (1996).
11. S. B. Feldman et al.,Virology 204, 1 (1994).
12. M. Feldman and P. Fitzgerald-Bocarsly, J. Interferon
Res. 10, 435 (1990).
13. J. J. Ferbas et al., J. Immunol. 152, 4649 (1994); S. B.
Feldman et al., J. Leukocyte Biol. 57, 214 (1995).
14. S. Ghanekar et al.,J. Immunol. 157, 4028 (1996).
15. C. Lopez, P. Fitzgerald, F. P. Siegal. J. Infect. Dis. 148,
962 (1983); D. M. Howell, S. B. Feldman, P. Kloser, P.
Fitzgerald-Bocarsly, Clin. Immunol. Immunopathol. 7,
223 (1994); J. Ferbas, J. Navratil, A. Logar, C. Rinaldo,
Clin. Diagn. Lab. Immunol. 2, 138 (1995).
16. F. P. Siegal et al.,J. Clin. Invest. 78, 115 (1986).
17. U. O’Doherty et al.,J. Exp. Med. 178, 1067 (1993).
18. U. O’Doherty et al.,Immunology 82, 487 (1994).
19. G. Grouard et al.,J. Exp. Med. 185, 1101 (1997).
20. J. Olweus et al.,Proc. Natl. Acad. Sci. U.S.A.94,
12551 (1997).
21. M.-C. Rissoan et al.,Science 283, 1183 (1999).
22.
Cells were incubated with UV-irradiated HSV in
quadruplicate wells (2 310
5
cells in 200 ml of culture
medium per well with 2 310
4
plaque-forming units of
virus in 96-well culture plates) (23). IFN in supernatants
from 24-hour cultures with and without IL-3 were ana-
lyzed with cytopathic reduction in human foreskin fibro-
blast monolayers cultured with vesicular stomatitis virus
(sensitivity, 2 to 25 IU of IFN per milliliter) (23). Cytocen-
trifuge preparations of cells from 6-hour cultures were
prepared for IFN-aimmunostaining with mouse mono-
clonal antibody 7N4-1 (10 mg/ml; Schering-Plough Re-
search Institute, Kenilworth, NJ) (21). Cells from 6-hour
cultures were centrifuged and the sediment frozen for
IFN-aand IFN-bmRNA PCR analyses (24).
23. F. P. Siegal et al.,Leukemia 8, 1474 (1994).
24. Reverse transcriptase (RT )–PCR: RNA was isolated
with the acid guanidinium thiocyanate–phenol-chlo-
roform method [P. Chomczrski and N. Sacchi, Anal.
Biochem. 162, 156 (1987)]. DNA contamination was
removed by digestion with deoxyribonuclease I (5 U;
Boehringer-Mannheim, Indianapolis, IN) for 30 min at
37°C. Controls without RT were performed. Reverse
transcription was carried out with pd( T)12-18 (Phar-
macia, Alameda, CA) priming and Superscript II RT
(Life Technologies, Grand Island, NY ). PCR reaction
volume was 25 ml containing 100 ng of each primer,
40 nM of each deoxynucleoside triphosphate, 1 mlof
cDNA, and 1.25 U of AmpliTaq (Perkin-Elmer, Foster
City, CA). The primers used were as follows: IFN-a
(sense: 59-GATGGCCGTGCTGGTGCTCA-39; antisense:
59-TGATTTCTGCTCTGACAACCTCCC-39; probe: 59-
CTCAAGCCATCTCTGTCCTCC ATGAGATGA-39);
IFN-b(sense: 59-TTGAATGGGAGGCTTGAATA-39;
antisense: 59-CTATGGTCCAGGCACAGTGA-39; probe:
59-GGCTGGAATGAGACTATTGTTGAGAACCTC-39);
and human ribosomal protein S14 (sense: 59-GGCA-
GACCGAGATGAATCCTCA-39; antisense: 59-CAGG-
TCCAGGGGTCTTGGTCC-39). Each PCR amplification
contained primers for the ribosomal protein S14, to
verify the amounts of cDNA. A GeneAmp PCR Sys-
tem 9700 (Perkin-Elmer/Applied Biosystems) was
used with an initial denaturation step of 94°C for 5
min, followed by cycles of 94°C for 30 s, 60°C for
30 s, 72°C for 60 s, and a final elongation step of
72°C for 7 min. PCR products were separated on a
2% agarose gel, followed by DNA blotting and hy-
bridization with
32
P-labeled probes.
25. S. Sun et al., J. Exp. Med. 188, 2335 (1998).
26. P. Marrack et al., ibid. 189, 521 (1999).
27. We thank T. L. Nagabhushan, L. L. Lanier, H. Kanzler,
D. Imperato, C. Lopez, L. Filgueira, G. Grouard, and R.
Rai for continued support. F.P.S. is supported by a
grant from Pharmacia and Upjohn. DNAX Research
Institute of Molecular and Cellular Biology is support-
ed by Schering-Plough Corporation.
8 February 1999; accepted 3 May 1999
Math1: An Essential Gene for
the Generation of Inner Ear
Hair Cells
Nessan A. Bermingham,
1,2
Bassem A. Hassan,
1,2
Steven D. Price,
7
Melissa A. Vollrath,
5
Nissim Ben-Arie,
2
* Ruth Anne Eatock,
5,6
Hugo J. Bellen,
1,2,4,5
Anna Lysakowski,
7
Huda Y. Zoghbi
1,2,3,4,5
†
The mammalian inner ear contains the cochlea and vestibular organs, which are
responsible for hearing and balance, respectively. The epithelia of these sensory
organs contain hair cells that function as mechanoreceptors to transduce sound
and head motion. The molecular mechanisms underlying hair cell development
and differentiation are poorly understood. Math1, a mouse homolog of the
Drosophila proneural gene atonal, is expressed in inner ear sensory epithelia.
Embryonic Math1-null mice failed to generate cochlear and vestibular hair cells.
This gene is thus required for the genesis of hair cells.
The inner ear initially forms as a thickening
of the ectoderm, termed the otic placode, be-
tween rhombomeres 5 and 6 in the hindbrain.
The otic placode gives rise to neurons of the
VIIIth cranial nerve and invaginates to be-
come the otocyst, from which the inner ear
Fig. 3. RT-PCR amplification of IFN-aand IFN-b. PCR products amplified from each cell population
after 20, 25, 30, or 35 cycles were separated on a 2% agarose gel containing ethidium bromide.
Negative controls contained no cDNA. Marker is 1-kb DNA Ladder (Life Technologies, Grand Island,
New York). IFN-amRNA is apparent in PBMCs, is increased in the DC2 precursors (pDC2s) with
enrichment and purification, and is diminished in the monocyte fraction. IFN-bmRNA is visualized
only in the most highly purified DC2 precursors.
REPORTS
www.sciencemag.org SCIENCE VOL 284 11 JUNE 1999 1837