Clin. exp. Immunol. (1986) 66, 77-87
A sensitive radioimmunoassay for alpha-interferon:
circulating a-interferon-like substance in the plasma
of healthy individuals and rheumatoid arthritis patients
S. SHIOZAWA, K. CHIHARA, K. SHIOZAWA, T. FUJITA, H. IKEGAMI, S.
KOYAMA & M. KURIMOTO Department of Medicine, Third Division, Kobe University
School of Medicine, Kobe, and Hayashibara Biochemical Laboratories, Okayama, Japan
(Acceptedfor publication 21 May 1986)
A radioimmunoassay for circulating a-interferon (IFNa) has been developed using
lymphoblastoid IFNa. The assay was specific for IFNa, and did not cross-react with
IFN#, IFNy, or ACTH, while it was specifically inhibited by recombinant IFNa. The
radioimmunoassay (y) correlated linearly with the virus inhibition assay (x), with a
regression line of y on x of y=0 659x+245 (u) (P<0 01). a-Interferon-like substance
(IFNa-LS) was extracted and concentrated from plasma either by silicic acid or by
antibody immunoadsorption. Serial dilutions of plasma and extracted samples of plasma
showed dilution curves identical to those of standard IFNa, suggesting the presence of
endogeneous IFNa in human plasma. The circulating IFNa-LS of healthy individuals
aged 20 to 45 was 0 207 + 0 055 ng/ml in males (n = 48) and 0 172 + 0-076 ng/ml in females
(n = 34). Gel filtration studies on a Sephadex G-75 column suggested that circulating
IFNat-LS exists in a fragmented form, inactive in virus inhibition assays, in the plasma of
healthy individuals. The finding may help explain why biological IFNa is often
undetectable in the plasma of healthy donors, yet is detectable by radioimmunoassay.
Circulating IFNa-LS in the plasma of healthy individuals declined gradually with age.
IFNot-LS was significantly decreased in the plasma of rheumatoid arthritis patients, when
compared with the value found in the age and sex-matched healthy controls and in
osteoarthritis patients (P < 0 0001). The decrease was related neither to treatment nor to
disease activity. IFNcx-LS was, however, not decreased in the plasma of vasculitis patients.
Decreased IFNa-LS in rheumatoid arthritis may be important from pathogenetic and
Keywords a-interferon radioimmunoassay rheumatoid arthritis ageing
IFNa is thought to play an important role in the immune response of the host (Svet-Moldavsky &
Chernyakhovskaya, 1967; Huang et al., 1971; Johnson & Baron, 1976; Schultz, Papamatheakis &
Chirigos, 1977; Skurkovich, Skorikova & Eremkina, 1978; Trinchieri & Santoli, 1978; Schultz,
Chirigos, & Heine, 1978; Herberman et al., 1980; Leanderson et al., 1982; Fradelizi & Gresser, 1982)
especially in the maintenance ofnatural killer cell activities (Svet-Moldavsky & Chernyakhovskaya,
1967; Skurkovich et al., 1978; Trinchieri & Santoli, 1978; Minato et al., 1980; Ortaldo et al., 1980;
Correspondence: S. Shiozawa, Department of Medicine, Third Division, Kobe University School of
Medicine, 7-5 Kusunokicho, Chuoku, Kobe 650, Japan.
78 S. Shiozawa et al.
Herberman et al., 1980; Targan & Dorey, 1980; Perussia & Trinchieri, 1981). However, in spite of
circumstantial evidence (Gresser et al., 1983; Buffet-Janvresse & Hovanessian, 1984), there has been
no definitive proof in support of the existence of endogenous IFNa. Circulating IFNa is often
undetected in the plasma of healthy individuals when measured by virus inhibition assays (Hooks et
al., 1979; 1982; Preble et al., 1982; Panem et al., 1982; Ytterberg & Schnitzer, 1982).
Rheumatoid arthritis is a chronic inflammation of the synovium, and extensive efforts to find
pathogenetic organisms in situ have so far been unrewarding. If rheumatoid arthritis is caused by
some unknown pathogenetic organism coming into the synovium through the systemic circulation,
as suggested by Zvaifler (1973), the ability of the host to handle invading organisms must have a
significant contribution to the sequelae of chronic inflammation. In this regard, IFNa may play an
important role, since it is necessary for continuous activation of natural killer cells, the cells
important for the host's front-line defence against invading organisms (Targan & Dorey, 1980;
Perussia & Trinchieri, 1981). In fact, aberrant natural killer cell activities and changes in circulating
IFNL% have been reported in rheumatoid arthritis (Goto et al., 1981; Neighbour & Grayzel, 1981;
Dobloug et al., 1982; Faure et al., 1983; Combe et al., 1984; Goto & Zvaifler, 1985). However, with
respect to the circulating level of IFNa, the results obtained using virus inhibition assays have been
inconsistent so far (Hooks et al., 1979; Degr6, Mellbye & Clarke-Jenssen, 1983).
In this study, we have measured circulating IFNa using a sensitive radioimmunoassay, based on
a polyclonal antibody made to lymphoblastoid IFNa as immunogen, which was specific for IFN2.
Circulating immunoreactive IFNx and ax-interferon-like substance (IFNa-LS), was extracted and
concentrated from human plasma, and the circulating level of IFNa-LS was determined in the
plasma of healthy individuals and rheumatoid arthritis patients. Gel filtration studies of IFNa-LS
suggested the existence of a fragmented form of IFNa in human plasma.
MATERIALS AND METHODS
Production ofIFNoe. Human leukaemic lymphoblastoid BALL- I cells (Miyoshi et al., 1977) were
grown up by subcutaneous transplantation into a new-born hamster which was pretreated with
anti-thymocyte antisera (Tanimoto, 1982). Tumour cells (5 x 106/ml) were then primed by 100 i.u./
ml of IFN for 1 h and cultured with 500 HA/ml of Sendai virus in RPMI 1640 medium at 370C for 24
h in a spinner vessel. The supernatant was harvested, acidified to pH 2 0 using 6 M HCl, and kept at
40C for at least 24 h to inactivate viruses.
Purification of IFNa. The culture supernatant, adjusted to pH 4-0 using 6 M NaOH, was applied
on a SP-Sephadex C-25 column (Pharmacia Fine Chemicals Inc., Uppsala, Sweden). The eluate
obtained by 0-1 M phosphate buffered saline, pH 8 0, was applied on a NK-2 Sepharose affinity
column for IFNa (Celltech Laboratories, Berkshire, England) (Allen et al., 1982). A single protein
peak was obtained by eluting with 0-1 M citrate buffer. The peak had specific activity of 5-4 x 108 i.U./
mg and 65% recovery. Further purification of the peak through Sephadex G-100 column
(Pharmacia Fine Chemicals) gave IFNa with specific activity of 9 5 x 107 i.u./mg and 53% recovery.
The final product showed a single band at mol. wt 17,000 + 1,200 daltons on SDS-polyacrylamide
gel electrophoresis with isoelectric point at pH 5-4. Its anti-viral activity was neutralized by anti-
IFNa (Mogensen, Pyhala & Cantell, 1975) but not by anti-IFNI sera (Mochida Pharmaceutical
Company, Tokyo, Japan) (Sidwell & Huffaman, 1971). The biological activities of purified IFNa,
such as species-specificity, inhibition of cellular multiplication, or activation of natural killer cells,
were similar to those of leukocyte IFNoc from human peripheral blood (Green Cross Company,
Osaka, Japan) (Matsuo, Hayashi & Kishida, 1974; Imanishi et al., 1980).
Virus inhibition assay. The antiviral activity of IFNac was assayed on FL cells as challenged by
Sindbis virus, and the result standardized according to the IFNa preparation MRC69/19B. The
result correlated well with an assay employing vesicular stomatitis virus. FL cells (5 x 105/ml) were
plated in a flat-bottomed multiwell microtitre plate (No. 163320, Nunc Ltd., Roskilde, Denmark)
and challenged by aliquots of sample or standard reference IFNa. After incubation for 24 h at 37°C,
Sindbis virus, optimally diluted to produce 50% cytotoxicity, was added to the reaction mixture,
and the culture continued for additional 24 h. Neutral red dye was added 45 min before harvesting,
Radioimmunoassay of circulating a-interferon 79
and the amount of dye taken up in the cell was measured under 540 nm photoabsorbance after
disrupting cells with 0 07 M glycine HCI, pH 3 0, and 30% ethanol (Pidot, 1971). In our hands, 25 i.u.
corresponded to 0 1 ng of IFNa protein.
Experimental inducation of IFNci in the peripheral blood. To induce high titres of IFN(i in the
human peripheral blood, 5 ml of heparinized blood were incubated with 500 HA/ml of Sendai virus
at 370C for various lengths of time less than 24 h, and the supernatant plasma used for assay.
Radioimmunoassay of IFNcx. Antibody to IFNa was raised in a rabbit by repeated subcutaneous
injection of IFNa with Freund's complete adjuvant, and was used for radioimmunoassay at a
dilution of 1:40,000. The purified IFNa was radio-iodinated by lactoperoxidase; the reaction
mixture contained 1-7 pg IFNcx, I mCi Na 1251 (New England Nuclear, Inc., Massachusetts, USA),
0 5 ug lactoperoxidase (Sigma Chemical Company, St Louis, USA), and 0 4 pg hydrogen peroxide
in 50 lp of 0 4 M sodium acetate buffer, pH 5 6. '25I-IFNa had a specific activity of approximately 0-3
mCi/pg. This was used for radioimmunoassay immediately after repurification on a Sephadex G-75
column (0 7 x 40 cm) (Pharmacia Fine Chemicals) using 0-07 M barbital buffer as eluent. The diluent
buffer for radioimmunoassay was 0-01 M phosphate buffer, pH 7 4, containing 0 16 M NaCl, 0.25%
bovine serum albumin, and 10% Triton X- 100. The assay was carried out in a test tube by adding 0 1
ml sample or standard IFNa, 0 1 ml anti-IFNax antisera, and 0 3 ml of diluent buffer. To measure
plasma samples, 0 1 ml of IFNax-free plasma was added in tubes containing serially diluted standard
IFNa. Human plasma was used after pretreatment with 20% (w/v) Norit A charcoal power to
absorb endogenous IFNa (IFNa-free plasma), since human plasma, but not sera from other species
of animals such as fetal bovine serum, significantly inhibited the binding of '25I-IFNa to antibody.
Separation of bound and free IFNa was carried out by adding goat anti-rabbit gammaglobulin sera.
The radioactivity of precipitates was counted using autogamma counter. The results were expressed
as bound per total radioactivities (B/T). Each samples were assayed in duplicate. The minimal
detectable quantity of IFNa, as defined by the amount causing 10% fall from the total replaceable
radioactivity (Bo), was approximately 6 pg/tube (0 06 ng/ml). Fetal bovine serum was purchased
from Microbiological Associates, Maryland, USA. For specificity control, recombinant IFNaX
(Ro22-8181, Roche Pharmaceuticals, New Jersey, USA), IFNJ (Kyowa Hakko Kogyo Co. Ltd,
Tokyo, Japan), IFNy (Interferon Sciences, Inc., New Jersey, USA), and ACTH were used.
Extraction of IFNa-LSfrom plasma. Silicic acid, 01 g, was added to 4 ml of plasma, and reacted
at room temperature for 1 h with stirring. The precipitate was washed with chilled distilled water
and extracted with 3 ml of 0 02 M HC1-80% acetone, and the supernatant was evaporated. After
reconstitution with the assay buffer and adjusting pH to 8 0, this was used for assay or filtered
through Sephadex G-75. To extract IFNat-LS by antibody immunoadsorption, gammaglobulin
fraction of anti-IFNat antisera was precipitated serially by 50% and 33% ammonium sulphate
solution. Affi-Gel 10 (Bio-Rad Laboratories, California, USA) was washed once with excess ice-
cold isopropanol followed by three washes with ice-cold distilled water. The gel slurry was
transferred to a plastic tube and briefly sedimented. Then approximately 2-5 ml bed volume of Affi-
Gel 10 was reacted with 50 mg of gammaglobulin fraction of anti-IFNaX antibody in 3-8 ml of 0 1 M
3-(N-morpholino) propane sulphonic acid (MOPS) for 1 h at room temperature then at 4°C
overnight. The gel was extensively washed with 0 1 M MOPS, and then reacted with plasma sample.
The reaction was carried out at 4°C for 4 h or 4°C overnight, in a polystyrene tube with continuous
rotation. The gel was washed with excess 0 01 M phosphate buffer, pH 7-4, containing 0 16 M NaCl
and 0 1% Triton X-100, centrifuged, and reacted with aliquots of 0 2 M acetic acid, pH 2 6,
containing 016 M NaCl and 01% Triton X-100 to extract IFNa. The supernatant was then
harvested and lyophilized.
Gel chromatography. Plasma samples, extracted either by silicic acid or antibody immunoad-
sorption, were dissolved in phosphate buffer, neutralized, and immediately applied on a Sephadex
G-75 column (0 9 x 44 cm) using 0-4 M ammonium acetate buffer, pH 4-0, as eluent. Elution was
carried out at 10 ml/h at room temperature, and each 1 ml fraction was collected using a peristaltic
pump system P- I (Pharmacia Fine Chemicals) and lyophilized. After reconstitution with the assay
buffer, each fraction was assayed for immunoreactive IFNa. Void volume (1VO) and bed volume (V,)
of the column were determined by blue dextran and saturated NaCl, respectively.
Protease inhibitors. To overcome proteolytic digestion of IFNa in plasma, various protease
80 S. Shiozawa et al.
inhibitors were used; the mixture of protease inhibitors used in this study contained 1 Yg/ml
pepstatin A (No. 4265, Sigma), 6 pg/ml leupeptin (Protein Research Foundation, Osaka, Japan), 20
pg/ml N-ethylmaleimide (Wako Chemicals), 200 pg/ml bacitracin (Wako), 5 mM EDTA (Wako), 20
pg/ml phenylmethylsulphonylfluoride (No. 7626, Sigma), and 2000 k.i.u./ml trazirol at final
concentration. The leupeptin solution was used at 250 pg/ml. To extract IFNa-LS from plasma,
blood was drawn with a syringe containing the mixture of protease inhibitors and immediately
centrifuged, and the supernatant plasma was stored at - 80'C until assayed.
Plasma and patient profile. Human blood sample was drawn, heparinized and immediately
centrifuged, and the plasma was stored at - 20'C or 80'C until assayed. Healthy donors without
apparent clinical signs of disease in the present and past were used as healthy individual. The patient
with systemic lupus erythematosus fulfilled the revised diagnostic criteria ofAmerican Rheumatism
Association (Tan et al., 1982).
Peripheral blood was obtained from 40 patients with rheumatoid arthritis (35 to 75 years old;
mean 50 1) including two males, and seven patients with osteoarthritis (48 to 64 years old; mean 54).
Rheumatoid arthritis patients fulfilled the diagnostic criteria of American Rheumatism Association
(Ropes et al., 1958). Rheumatoid patients were considered active if the patient exhibited all of the
following: (1) arthritis with pain and/or swelling in more than four joints by physician's assessment,
(2) increased tendency to fatigue and more than 1 h of morning stiffness, (3) increased erythrocyte
sedimentation rate exceeding 40 mm for the first hour and/or increased C-reactive protein. Patients
were considered to be inactive by applying the criteria for clinical remission (Pinals et al., 1982).
Patients with rheumatoid arthritis were either without treatment or had 25 mg gold thiomalate
injection biweekly, plus a daily non-steroidal anti-inflammatory drug equivalent to 25-50 mg of
indomethacin (shown in Fig. 6). Plasma samples of 49 rheumatoid patients (mean 45 4 years old) at
F 0 -10 _
50 P 8E -30I
0 100 1000 10000
0\ /39- interferon (iu/ml)
10~ ~ ~ N* 1
.00k < 100 1000 10000
y- inteferon ( u/ml)
v 'i 0.1 0 4 2 4
I I I I
25 100 250 1000
Fig. 1. Standard curve of a-interferon radioimmunoassay. The vertical axis gives the percentage o1 precipitated
versus total 125I-a-interferon (%B/T). The horizontal axis gives the amount of serially diluted standard a-
interferon. Human leukocyte BALL-1 a-interferon plasma added (0); BALL-1 a-interferon without plasma
(0); recombinant a-interferon (0). Inhibition curves with fl-interferon or y-interferon are shown in panel A and
B, respectively. Human plasma, added to standard dilution tubes, is preabsorbed with 20% (w/v) charcoal
powder to absorb endogenous a-interferon. a-Interferon of recombinant DNA origin, Ro 22-8181, shows a
similar displacement curve to BALL- 1 cell a-interferon. 0-1 ng of a-interferon protein corresponds to 25 iu of
virus inhibition activity.
Radioimmunoassay of circulating a-interferon 8i
the North Hyogo Medical and Orthopedic Center, drawn with EDTA, were also examined in this
study. In this group of patients, approximately 80% were receiving 25 mg of gold thiomalate
injection twice weekly, 30% had 125 mg D-penicillamine daily. Almost all received non-steroidal
anti-inflammatory drugs, equivalent to 25-50 mg ofindomethacin daily. Rheumatoid vasculitis was
diagnosed on a clinical and/or histological basis; these included the patients with cutaneous
vasculitis and/or interstitial pneumonitis (Cupps & Fauci, 1981). Patients with osteoarthritis
receiving non-steroidal anti-inflammatory drugs equivalent to 50 mg of indomethacin were also
Statistics. For calculating mean + s.d., the value of IFNa-LS less than 10% fall of sensitivity was
estimated to be 0-05 ng/ml, based on the finding that, when plasma IFNx-LS of rheumatoid arthritis
patient was concentrated and measured, most of the plasma level were in approximately 0 05 ng/ml
(not shown). Statistical comparison was carried out using Student's t-test modified for a small
sample (Hoel, 1966).
Specificity of radioimmunoassay. The standard curve of IFNa radioimmunoassay in buffer and
also in human plasma is shown in Fig. 1. Human plasma, but not sera from other species of animals
such as fetal bovine serum, significantly inhibited the binding of '251-IFNa to antibody, when added
without pretreatment with charcoal, suggesting that endogenous IFNa may be present in human
plasma. Charcoal-treated plasma however did not interfere with standard dilution curve (Fig. 1).
The specificity of the radioimmunoassay was confirmed because (a) the assay did not cross-react
with IFNf, IFNy, or with up to 0 34 pg of ACTH; and (b) recombinant IFN7x Ro22-8 181 inhibited
the binding of '251-IFNa to a comparable level as BALL-l cell IFNcx (Fig. 1).
Bioassay and characterization of IFNa-LS. Linear correlation existed between the radioimmu-
noassay (y) and virus inhibition assay (x), with a regression line of y on x as y = 0 659x+ 245 (u)
(P < 0-01) (Fig. 2). Intra- and inter-assay coefficients of variation of radioimmunoassay were 0-2 1%
and 4-12%, respectively. The minimal detectable quantity of IFNa was approximately 0-06 ng/ml (6
pg/tube). IFNa-LS was reproducibly measured in plasma after two freeze-thaw cycles. Serial
dilutions of plasma of representative individuals well coincided with the standard curve (not
. 20 5000
Xo / ~~~~~
qo / _ ~~~~~3000_
0 5000 10000
Bioassay of a-interferon (iu/ml )
Fig. 2. Correlation between radioimmunoassay and virus inhibition assay. Heparinized whole blood (5 ml) is
incubated with 500 HA/ml ofSendai virus at 37°C for a variable time less than 24 h, and the supernatant plasma
is assayed simultaneously by radioimmunoassay and virus inhibition assay. The vertical axis (y) gives protein
determination by radioimmunoassay (ng/ml), which is correspondingly translated into equivalent amounts of
virus inhibition activity (i.u./ml), based on the correspondence of 25IU to 0 1 ng of a-interferon protein. The
horizontal axis (x) gives the value of virus inhibition assay (i.u./ml).
82 S. Shiozawa et al.
shown). The IFNa-LS was extracted from plasma by silicic acid, with approximately 20% yield.
And the serial dilutions of the extracted sample of plasma also exactly coincided with the standard
curve. Circulating IFNa-LS was also extractable by anti-IFNa antibody immunoadsorption, with
approximately 90% yield. This type of extraction experiments were reproducibly repeated many
times for gel filtration analysis.
Gelfiltration of IFNc. To further characterize the circulating IFNa-LS, plasma from healthy
donors was extracted by antibody immunoadsorption and filtered through a Sephadex G-75
column, and each fraction assayed for immunoreactive IFNa. Unexpectedly IFNa-LS eluted in a
position of distinctively smaller molecular size, corresponding to fraction 16 in Fig. 3, when
compared with that of standard IFNox (not shown). To rule out the possibility of artefactual
degradation of IFNa during the extraction procedure, IFNa of standard molecular size was added
to plasma and immediately extracted using anti-IFNa antibody-coupled Affi-GellO, with or
without protease inhibitors (Fig. 3). Extraction was carried out at 40C overnight. Figure 3a shows
the elution profile of IFNa filtered through a Sephadex G-75 column. While IFNa without protease
inhibitors eluted in fractions 15-17, IFNa treated with the mixture of protease inhibitors (See
Materials and Methods) eluted into the authentic peak of IFNa at fractions 9-11. Figure 3b shows
an elution profile of IFNa treated likewise by adding 250 yg/ml of leupeptin. The addition of
leupeptin to the reaction mixture effectively prevented the shift of immunoreactive peak of IFNa,
suggesting that fragmentation of IFNa occured in plasma during the process of affinity extraction.
Gel filtration of circulating IFNot-LS. Based on the above findings, plasma from health
individuals was re-examined in the presence of a fully inhibitory dose of protease inhibitors to avoid
artefactual degradation during extraction procedure (Fig. 3c). Plasma was drawn with a syringe
containing the mixture of protease inhibitors and immediately processed for affinity extraction by
adding protease inhibitors in an identical manner to Fig. 3a. Figure 3c shows a gel filtration profile
of plasma oftwo representative healthy donors. It can be seen that the majority of circulating IFNa-
LS eluted in the second peak at fractions 15-17, while a minute peak remained in fractions 9- 11
(Fig. 3c), suggesting that plasma of healthy individuals contained a predominantly fragmented
form of IFNo-LS. In a preliminary experiment, plasma of an active lupus patient was filtered
through Sephadex G-75 and assayed for immunoreactive IFNa. Even though plasma was extracted
in the absence of protease inhibitors, a substantial part of the immunoreactive IFNa remained in the
authentic peak at fractions 9-11 (Fig. 3d), suggesting that the plasma of an active lupus patient
contained a certain amount of intact molecules of IFNa. This may explain why biologically
detectable IFNa was present in the lupus patient's plasma (Hooks et al., 1979; 1982; Preble et al.,
1982; Panem et al., 1982; Ytterberg & Schnitzer, 1982). We were unable to examine the plasma of
the lupus patient in the presence of protease inhibitor.
Quantification of IFNa-LS in the plasma of healthy individuals. Circulating IFNa-LS in the
plasma of healthy individuals at the age 20-45 was 0207+00055 ng/ml in 48 males and
0-172+0 076 ng/ml in 34 females (Fig. 4). There were minor monthly variations in the level of
circulation IFNa-LS, suggesting that there may be some seasonal variation of circulation IFNa-LS
since all the assays were standardized using an internal standard; this preliminary observation
requires confirmation. IFNa was not detectable, however, in the same sample of plasma when
assayed by virus inhibition assays (not shown). In our hands, the minimal detectable quantity of
IFNa in virus inhibition assays was 2 i.u./ml, which corresponds to 0-008 ng/ml of IFNa protein.
Thus, when IFNa-LS was intact and biologically active, a circulating IFNae level of approximately
0-2 ng/ml should be detectable under virus inhibition assays.
It was found that circulating IFNac-LS gradually declined with age (Fig. 5).
Quantification of IFNa-LS in rheumatoid arthritis. Circulating IFNa-LS was significantly
decreased in the plasma of rheumatoid arthritis patients when compared with the value found in the
age and sex-matched healthy control (P< 00001) (Fig. 6). It was decreased irrespective of the
disease activity. Circulating IFNcx-LS was also decreased in the active patients not treated for
rheumatoid arthritis (P< 0001), whereas it was not decreased in the rheumatoid patients with
vasculitis. Since rheumatoid patients commonly received high dose of non-steroidal anti-
inflammatory drugs, it was necessary to examine the contribution of the drugs to circulating IFNa-
LS. Circulating IFNx-LS in osteoarthritis patients receiving equivalent dose of the non-steroidal
Radioimmunoassay of circulating tx-interferon 83
5 10 15 20
Fig. 3. Sephadex G-75 column chromatography (0 9 x 44 cm). Elution is carried out using 0 1 M ammonium
acetate buffer, pH 4 0, at room temperature at a rate 10 ml/h. Each 1 ml is fractionated, lyophilized, and
reconstituted with the assay buffer. (a) Standard a-interferon is added to plasma and immediately processed for
extraction using antibody immunoadsorption. Extraction is carried out at 40C overnight with rotation. The
mixture of protease inhibitors (see Materials and Methods) is present (a A) or absent (---- -A). (b)
Standard a-interferon is added to plasma and extracted as in A, in the presence (0 O) or absence (-----0)
of 250 ,ug/ml of leupeptin. (c) Plasma is drawn and affinity extracted from two healthy individuals in the presence
of the mixture of protease inhibitors in a manner identical to A. (d) Plasma of an active patient with systemic
lupus erythematosus is drawn and extracted without protease inhibitors (v v). The elution peak of standard
a-interferon is shown ( ....... ). Each point represents the mean + s.e.m. of determinations performed in
anti-inflammatory drugs was not decreased when compared with the age and sex-matched healthy
control (Fig. 6.). Plasma ofanother 49 rheumatoid arthritis patients from the North Hyogo Medical
& Orthpaedic Center, drawn with EDTA, showed similar low levels of circulating IFNa-LS
0-069 + 0049 ng/ml, a significant decrease when compared with that of age-matched healthy
control, 0 120 + 0050 ng/ml (P < 0005); circulating IFNa-LS was below the sensitivity limit of
assay in 39 out of 49 patients examined.
We have detected circulating IFNa-LS in the plasma of healthy individuals using a sensitive
radioimmunoassay. Since the radioimmunoassay is a specific for a-type of IFN, recognizing
84 S. Shiozawa et al.
05 r (a) ( b)
Mean plasma level
Mar Apr May Jun Male Female
Fig. 4. (a) Monthly variation of circulating a-interferon-like substance in plasma of healthy individuals. Data
from the same donor is connected with a line. (b) Circulating a-interferon-like substance in the plasma of healthy
individuals aged between 20 and 45. Each point represents the mean of determinations performed in duplicate.
0-3 0 a
00 000 I 0 o
20 30 50 60-85
Fig. 5. Age distribution of circulating a-interferon-like substance in plasma of healthy donors. Each point
represents the mean of determinations performed in duplicates. Values in each age group are given as the
mean + s.e.m. Statistical significance of difference between each age group of all combinations is as follows; in
males, age 30-40; P< 00001, 30-50; P < 0 0001, 30-60; P < 0-0001, 20-40; P < 0 05, 20-50; P< 0 00025, 20-60;
P<001, and otherwise not significant. In females, 30-40; P<0005, 30-50; P<0005, and otherwise not
Radioimmunoassay of circulating ae-interferon 85
P<OOOO P<O-OOO NS NS
RA RA RA OA
active inactive vasculitis
Fig. 6. Radioimmunoassay of circulating a-interferon-like substance in the plasma of patients with rheumatoid
arthritis (RA). Each point represents the mean of determinations performed in duplicates. The values given are
the mean + s.e.m. The shaded area represents the mean + s.e.m. of age and sex-matched healthy controls. Patient
without drugs (0); patient treated with non-steroidal anti-inflammatory drug plus 25 mg of gold thiomalate
injection biweekly (0). osteoarthritis (OA). Statistical significance by Student's t-test is given.
biologically active as well as inative IFNa, it is unlikely that the assay recognized antigenically
similar but unrelated peptide in plasma crossreacting with IFNa. This is supported from the
following evidence: (1) serial dilutions of plasma gave identical dilution curves to that of standard
IFNa; (2) circulating IFNa-LS was extracted and concentrated from plasma under acidic condition
using silicic acid and HCl as well as under neutral condition using antibody immunoadsorption; (3)
extracted plasma samples showed dilution curves identical to that of standard IFNa. Definitive
proof for this, however, awaits the determination of the amino acid sequence of IFNaC-LS.
Gel filtration studies suggested that circulating IFNa-LS appeared to be fragmented in human
plasma; (1) circulating IFNa-LS of the healthy individuals eluted into the fraction of distinctively
smaller molecular size than that of standard IFNax; ( 2) when standard IFNa was added to plasma
and extracted without adding protease inhibitors, immunoreactive peak of IFNa moved from its
original position to that of smaller molecular size, whereas (3) the shift was preventable by adding
protease inhibitors. Since this shift was prevented by leupeptin, plasmin-like factors (Mantei et al.,
1980; White & Gross, 1957) might be responsible for degradation. The existence of a fragmented
and biologically inactive form of IFNax in plasma is consistent with the finding that IFNa is often
undetectable in the plasma of healthy individuals when assayed by virus inhibition assays are shown
in this and other studies (Hooks et al., 1979; 1982; Preble et al., 1982; Panem et al., 1982; Ytterberg
& Schnitzer, 1982). Theoretically, since the plasma of a healthy individual contained approximately
0 2 ng/ml of IFNa-LS (corresponding to 50 i.u./ml of antiviral activity), it should be detectable
under virus inhibition assays when the IFNa molecule was intact and biologically active.
It was found in this study that circulating IFNa-LS declined gradually with age. This may be a
part of the gradual fall in immune surveillance in the aged, since IFNa is important for activation of
natural killer cells (Targan & Dorey, 1980; Perussia & Trinchieri, 1981).
The results also showed that circulating IFNax-LS was significantly decreased in the plasma of
rheumatoid arthritis patients. This is attributable to the original disease condition, as suggested by
the following evidence; (1) circulating IFNat-LS was decreased in the patients untreated for
rheumatoid arthritis; (2) it was decreased when compared with the value found in the osteoarthritis
86 S. Shiozawa et al.
patients receiving a comparable dose of non-steroidal anti-inflammatory drugs; (3) plasma sample
drawn with either heparin or EDTA gave similar results, indicating that the result was irrelevant to
the method of sampling; (4) IFNa-LS was however not decreased in the rheumatoid patients with
vasculitis. The latter finding might be interesting in relation to the vascular reactivity of IFNoa.
Kramer et al. (1984) have shown that IFNx potentiated acute vascular rejection in the renal
'transplant recipients; IFNa appears to be related to the vascular reactivity.
There have been inconsistency in results concerning the circulating level of IFNa in the
rheumatoid arthritis patients; previous assays based on antiviral activities of IFNa have shown that
circulating IFNa was increased in rheumatoid arthritis (Hooks et al., 1979), whereas it was later
suggested to be IFNy (Degre et al., 1983). This type of uncertainty would be inevitable when the
virus inhibition assays were used (Hooks et al., 1979; Preble et al., 1982). The present results
utilizing radioimmunoassay seems to answer this by demonstrating that circulating IFNoc-LS is
significantly decreased in rheumatoid arthritis. Decreased IFNox in rheumatoid arthritis may be
important from the pathogenetic as well as therapeutic standpoints. Decreased IFNa in these
patients may lead to defective functions of natural killer cells (Goto et al., 1981; Combe et al., 1984;
Goto & Zvaifler, 1985), then prolong chronic inflammation, since IFNa is required for continuous
activation of natural killer cells (Targan & Dorey, 1980; Perussia & Trinchieri, 1981). In-vitro
production of IFNa by the patient's lymphocytes was decreased in rheumatoid arthritis (Neighbour
& Grayzel, 1981). The finding of decreased IFNcx-LS may also provide a rational basis for the
treatment of rheumatoid arthritis with IFNa; we have noted that, in various disease conditions,
clinical efficacy of IFNox treatment appeared to be related to the level of circulating IFNcx (to be
This work was supported in part by the Basic Researches for Joint Diseases Award and the Japan Ministry of
Education & Welfare grant 59770445 to S.S. We thank colleagues in the North Hyogo Medical and Orthopaedic
Center and Drs Ikuo Kobayashi, Osamu Ohno, and Yasuhiko Okimura for continuing help.
ALLEN, G., FANTES, K.H., BURKE, D.C. & MORSER, J. FRADELIZI, D. & GRESSER, I. (1982) Interferon inhibits
(1982) Analysis and purification of human lym- the generation of allospecific suppressor T lympho-
phoblastoid (Namalwa) interferon using a mono- cytes. J. exp. Med. 155, 1610.
clonal antibody. J. gen. Virol. 63, 207. GOTO, M., TANIMOTO, K., CHIHARA, T. & HORIUCHI,
BUFFET-JANVRESSE, C. & HOVANESSIAN, A.G. (1984) Y. (1981) Natural cell-mediated cytotoxicity in
Enzyme markers for the presence of circulating Sj0gren's syndrome and rheumatoid arthritis. Arth-
interferon: 2-SA synthetase in blood lymphocytes ritis Rheum. 24, 1377.
and protein kinase in platelet-rich plasma (41783). GOTO, M. & ZVAIFLER, N.J. (1985) Impaired killer cell
Proc. Soc. exp. Biol. Med. 175, 169. generation in the autologous mixed leukocyte reac-
COMBE, B., POPE, R., DARNELL, B. & TALAL, N. (1984) tion by rheumatoid arthritis lymphocytes. Arthritis
Modulation of natural killer cell activity in the Rheum. 28, 731.
rheumatoid joint and peripheral blood. Scand. J. GRESSER, I., BELARDELLI, F., MAURY, C., MAUNOURY,
Immunol. 20, 551. M-T. & TOVEY, M.G. (1983) Injection of mice with
Cupps, T.R. & FAUCI, A.S. (1981) The Vasculitides. antibody to interferon enhances the growth of
p. 1. W.B. Saunders, Co., Philadelphia. transplantable murine tumors. J. exp. Med. 158,
DEGRt, M., MELLBYE, O.J. & CLARKE-JENSSEN, 0. 2095.
(1983) Immune interferon in serum and synovial HERBERMAN, R.B., ORTALDO, J.R., DJEU, J.Y., HOL-
fluid in rheumatoid arthritis and related disorders. DEN, H.T., JETT, J., LANG, N.P., RUBINSTEIN, M. &
Ann. rheum. Dis. 42, 672. PESTKA, S. (1980) Role of interferon in regulation of
DOBLOUG, J.H., F0RRE, 0., KVIEN, T.K., EGELAND, T. cytotoxicity by natural killer cells and macro-
& DEGRt, M. (1982) Natural killer (NK) cell phages. Ann. N.Y. Acad. Sci. 350, 63.
activity of peripheral blood synovial fluid, and HOEL, P.G. (1966) Elementary Statistics. John Wiley
synovial tissue lymphocytes from patients with & Sons Inc., New York.
rheumatoid arthritis and juvenile rheumatoid arth- HOOKS, J.J., MOUTSOPOULOS, H.M., GEIS, S.A.,
ritis. Ann. rheum. Dis. 41, 490. STAHL, N.I., DECKER, J.L. & NOTKINS, A.L. (1979)
FAURE, G., BENE, M.C., TAMISIER, J.N. & THOMAS, P. Immune interferon in the circulation of patients
(1983) Monoclonal antibody (HNK1-Leu7) with autoimmune disease. New Engl. J. Med. 301, 5.
defined lymphoid cells in the blood of rheumatoid HOOKS, J.J., JORDAN, G.W., Cupps, T., MOUTSOPOU-
arthritis patients. Arthritis Rheum. 26, 1173. LOS, H.M., FAUCI, A.S. & NOTKINs, A.L. (1982)
Radioimmunoassay of circulating a-interferon 87
Multiple interferons in the circulation of patients PINALS, R.S., BAUM, J., BLAND, J., FOSDICK, W.M.,
with systemic lupus erythematosus and vasculitis. KAPLAN, S.B., MASI, A.T., MITCHELL, D.M.,
Arthritis Rheum. 25, 396. ROPES, M.W., SHORT, C.L., SIGLER, J.W. & WEIN-
HUANG, K-Y., DONAHOE, R.M., GORDON, F.B. & BERGER, H.J. (1982) Preliminary criteria for clinical
DRESSLER, H.R. (1971) Enhancement of phago- remission in rheumatoid arthritis. Bull. Rheum. Dis.
cytosis by interferon-containing preparations. 32, 7.
Infect. Immunity 4, 581. PREBLE, O.T., BLACK, R.J., FRIEDMAN, R.M., KLIP-
IMANISHI, J., PAK, C-B., KAWAMURA, H., KITA, M., PEL, J.H. & VILCtK, J. (1982) Systemic lupus
SUGINO, S., SAKAMOTO, M., TANIMOTO, T., erythematosus: presence in human serum of an
MASUDA, K., YOKOBAYASHI, K., MITSUHASHI, M., unusual acid-labile leukocyte interferon. Science
NAGANO, Y. & KISHIDA, T. (1980) Production and 216, 429.
characterization of interferon from human leuke- RoPES, M.W., BENNETT, G.A., COBB, S., JACOX, R. &
mic lymphoblastoid cells grown in hamsters. J. JESSAR, R.A. (1958) 1958 revision of diagnostic
appl. Biochem. 2, 257. criteria for rheumatoid arthritis. Bull. Rheum. Dis.
JOHNSON, H.M. & BARON, S. (1976) The nature of the 9, 175.
suppressive effect of interferon and interferon SCHULTZ, R.M., PAPAMATHEAKIS, J.D. & CHIRIGOS,
inducers on in vitro immune response. Cell. Immu- M.A. (1977) Interferon: an inducer of macrophage
nol. 25, 106. activation by polyanions. Science 197, 674.
KRAMER, P., TEN KATE, F.W.J., BIJNEN, A.B., JEEKEL, SCHULTZ, R.M., CHIRIGOS, M.A. & HEINE, U.I. (1978)
J. & WEIMAR, W. (1984) Recombinant leucocyte Functional and morphologic characteristics of
interferon A induces steroid-resistant acute vascu- interferon-treated-macrophages. Cell. Immunol. 35,
lar rejection episodes in renal transplant recipients. 84.
Lancet i, 989. SIDWELL, R.W. & HUFFMAN, J.H. (1971) Use of
LEANDERSON, T., HILL6RN, V., HOLMBERG, D., LARS- disposable micro tissue culture plates for antiviral
SON, E-L. & LUNDGREN, E. (1982) Selective effects and interferon induction studies. Appl. Microbiol.
on interferon on distinct sites of the T lymphocyte 22, 797.
triggering process. J. Immunol. 129, 490. SKURKOVICH, S.V., SKORIKOVA, A.S. & EREMKINA,
MANTEI, N., SCHWARZSTEIN, M., STREULI, M., E.I. (1978) Enhancement by interferon of lympho-
PANEM, S., NAGATA, S. & WEISSMANN, C. (1980) The cyte toxicity in normal individuals to cells ofhuman
nucleotide sequence of a cloned human leukocyte lymphoblastoid lines. J. Immunol. 121, 1173.
interferon cDNA. Gene 10, 1. SVET-MOLDAVSKY, G.J. & CHERNYAKHOVSKAYA, I.JU.
MATSUO, A., HAYASHI, S. & KISHIDA, T. (1974) (1967) Interferon and the interaction of allogeneic
Production and purification of human leukocyte normal and immune lymphocytes with L cells.
interferon. Japan J. Microbiol. 18, 21. Nature (Lond.) 215, 1299.
MINATO, N., REID, L., CANTOR, H., LENGYEL, P. & TAN, E.M., COHEN, A.S., FRIES, J.F., MASI, A.T.,
BLOOM, B.R. (1980) Mode of regulation of natural MCSHANE, D.J. ROTHFIELD, N.F., SCHALLER, J.G.,
killer cell activity by interferon. J. exp. Med. 152, TALAL, N. & WINCHESTER, R.J. (1982) The 1982
124. revised criteria for the classification of systemic
MIYosHI, I., HIRAKI, S., TSUBOTA, T., KUBONISHI, I., lupus erythematosus. Arthritis Rheum. 25, 1271.
MATSUDA, Y., NAKAYAMA, T., KISHIMOTO, H. & TANIMOTO, T. (1982) Production and purification of
KIMURA, I. (1977) Human B cell, T cell and null cell interferon from human leukemic lymphoblastoid
leukemic cell lines derived from acute lymphoblas- cells (BALL-1) grown in hamster. J. Kyoto pref.
tic leukemias. Nature (Lond.) 267, 843. Univ. Med. 91,1321.
MOGENSEN, K.E., PYHALA, L. & CANTELL, K. (1975) TARGAN, S. & DoREY, F. (1980) Dual mechanism of
Raising antibodies to human leukocyte interferon. interferon augmentation of natural killer cytotoxi-
Acta Path. Microbiol. Scand. Sect. B. 83, 443. city (NKCC). Ann. N. Y. Acad. Sci. 350, 121.
NEIGHBOUR, P.A. & GRAYZEL, A.I. (1981) Interferon TRINCHIERI, G. & SANTOLI, D. (1978) Antiviral acti-
production in vitro by leukocytes from patients vity induced by culturing lymphocytes with tumor-
with systemic lupus erythematosus and rheumatoid derived or virus-transformed cells. Enhancement of
arthritis. Clin. exp. Immunol. 45, 576. human natural killer cell activity by interferon and
ORTALDO, J.R., PHILLIPS, W., WASSERMAN, K., & antagonistic inhibition of susceptibility of target
HERBERMAN, R.B. (1980) Effects of metabolic in- cells to lysis. J. exp. Med. 147, 1314.
hibitors on spontaneous and interferon-boosted WHITE, W.F. & GROSS, A.M. (1957) Adrenocortico-
human natural killer cell activity. J. Immunol. 125, tropin. XIV. Action of bovine fibrinolysin and of
1839. liver cathepsin on corticotropin A: effect on biologi-
PANEM, S., CHECK, I.J., HENRIKSEN, D. & VILCEK, J. cal activity. J. Am. Chem. Soc. 79, 1141.
(1982) Antibodies to a-interferon in a patient with YTTERBERG, S.R. & SCHNITZER, T.J. (1982) Serum
systemic lupus erythematosus. J. Immunol. 129, 1. interferon levels in patients with systemic lupus
PERUSSIA, B. & TRINCHIERI, G. (1981) Inactivation of erythematosus. Arthritis Rheum. 25, 401.
natural killer cell cytotoxic activity after interaction ZVAIFLER, N.J. (1973) The immunopathology ofjoint
with target cells. J. Immunol. 126, 754. inflammation in rheumatoid arthritis. Adv. Immu-
PIDOT, A.L.R. (1971) Dye uptake assay: an efficient nol. 16, 265.
and sensitive method for human interferon
titration. Appl. Microbiol. 22, 671.