Molecular Brain Research, 14 (1992) 293-301 © 1992 Elsevier Science Publishers B.V. All rights reserved. 0169-328X/92/$05.00

293

BRESM 70446

The effect of active serum albumin on PC12 cells" I. Neurite retraction and activation of the phosphoinositide second messenger system David Dyer, Gabor Tigyi and Ricardo Miledi Laboratory of Cellular and Molecular Neurobiology, Department of Psychobiology, University of California, Irvine, Irvine, CA 92717 (USA) (Accepted 10 March 1992)

Key words: PC12 cell; Neurite; Serum albumin; Inositol trisphosphate; Retraction

Vertebrate blood sera contain a factor that triggers oscillatory chloride currents in Xenopus oocytes through activation of the phosphoinositide/Ca2÷ second messenger system. The active serum component consists of lipids bound to an isoform of serum albumin that we have named active serum albumin (ASA). In undifferentiated PC12 cells, micromolar concentrations of ASA inhibit the early morphological changes induced by NGF, whereas in differentiated PC12 cells ASA caused a rapid withdrawal of neurites, which was reversible and dependent upon culture age. In contrast to normal serum, plasma and thrombin did not cause neurite retraction. Preincubation of ASA with monospecific antibodies to serum albumin suppressed its ability to induce neurite retraction in a dose dependent fashion. As in the oocyte, ASA activated the phosphatidylinositol second messenger system of PC12 cells, causing a several fold increase in Ins1,4,5P3 levels within minutes of application. The Insl,4,5P3 increase was also blocked, in a titratable fashion, when ASA was preincubated with monospecific antibodies to serum albumin. This suggests that ASA-induced neurite retraction in PC12 ceils may depend, at least in part, on activation of the phosphatidylinositol second messenger system. Results involving albumin-depleted sera show that ASA is the main factor responsible for serum vulnerability of neurites in PC12 cells. These findings point to some limitations in the use of serum in culture media, and raise the possibility that the serum factor may impair neuronal plasticity in disorders that are accompanied by the activation of blood coagulation together with a breakdown of the blood-brain barrier. INTRODUCTION Serum is one of the richest sources of nutrients and factors required for cell growth and is, therefore, widely used as a m e d i u m supplement for the in vitro culture of m a m m a l i a n cells, including neurons. In spite of this, there are valid reasons to question the use of serum for in vitro n e u r o n a l culture. For example, cerebrospinal fluid is different from both serum and plasma, having a protein concentration 200-fold lower than that of serum 13. Moreover, serum is formed from plasma in the course of blood coagulation, and is found within the central nervous system only during abnormal conditions associated with breakdown of the b l o o d - b r a i n barrier and activation of blood clotting. Therefore, culture media which use serum as a supplement do not faithfully mimic the normal e n v i r o n m e n t of neurons, but instead create a milieu that is e n c o u n t e r e d only in pathological conditions. F u r t h e r m o r e , serum has b e e n shown to cause the withdrawal of neurites in differentiated PC12 cells in culture 24'25, and has also been reported to inhibit neurite outgrowth in other n e u r o n a l culture systems 1'9'3°'31.

Differentiating n e u r o n s extend processes and establish synapses with other cells in response to various growth factors in their environment. A good deal of the present knowledge about n e u r o n a l differentiation comes from studies of nerve growth factor (NGF) which promotes neurite outgrowth in sympathetic neurons and rat pheochromocytoma PC12 cells 14'15'19. In contrast, very little is known about factors which inhibit neurite outgrowth 8,16. We recently discovered a serum factor that elicits oscillatory chloride currents by activating the phosphoinositide (PI) second messenger system in Xenopus oocytes 26'27'28. This factor, which we have provisionally n a m e d active serum a l b u m i n (ASA), is a lipid modification of albumin produced during blood coagulation26'28. In search of the physiological function of this novel serum factor, we have tested its effect on a variety of m a m m a l i a n cell lines cultured in vitro in serum-free, chemically defined media. Some of our findings have already been presented in preliminary form elsewhere 1°' 11,26. Here we report our findings on the effects of the serum factor on nerve growth factor-induced differenti-

Correspondence: D. Dyer, Department of Psychobiology, UCI, Irvine, CA 92717, USA. Fax: (i) (714) 725-2447.

294 a t i o n in P C 1 2 cells, w h i l e its e f f e c t s o n cell p r o l i f e r a t i o n will b e p r e s e n t e d e l s e w h e r e in d e t a i l . MATERIALS AND M E T H O D S

2 nM in DMEM-N1. Typically, after 2 days of differentiation, approximately 75-80% of the cells developed neurites which were 15-30 ftm in length. Morphological changes in cultures exposed to ASA were monitored by time-lapse phase contrast photomicrography, by placing the microscope in a humidified 37°C incubator (5% CO_,).

Preparation of serum and plasma Human plasma and serum were used because of their obvious medical pertinence, and were obtained from whole blood drawn from healthy donors into sterile Vacutainer tubes (Beckton Dickinson, Rutherford, N J). Plasma samples were prepared in the presence of either heparin (18.1 U/ml) or hirudin (5 U/ml), followed by centrifugation for 15 min at 3000 x g to separate plasma. Plasma samples were frozen in liquid nitrogen immediately after centrifugation and thawed less than 2 rain before addition to cultures. Serum was prepared by allowing blood to coagulate for 1 h at 37°C, incubating at 4°C overnight. Serum was then poured off of the clot and centrifuged as described above to remove blood cells. Fresh plasma samples showed no activity in the oocyte assay at dilutions as low as 1:10, whereas serum samples were active at dilutions as high as 1:10 5. As a control for heparin present in some of the plasma samples, heparin was added to some serum samples at a concentration of 18.1 U/ml.

Purification of ASA ASA was purified from human serum according to Tigyi et al. 2~. Briefly, albumin was isolated from serum using Cibacron Blue-F3-6-A Agarose affinity chromatography. The bound albumin fraction was further purified by Concanavalin A-Sepharose affinity chromatography, followed by the separation of A S A and inactive albumin by hydroxyapatite chromatography. Both A S A and inactive albumin fractions were electrophoretically homogenous, comigrated with the human serum albumin (HSA) standard, and had identical N-terminal amino acid sequences. ASA activity was monitored by the albumin fractions' ability to elicit oscillatory chloride currents in Xenopus oocytes 27. In most experiments, different lots of commercially available HSA (fraction V: >96% pure; e.g. Sigma catalogue no. A-1653) were used, as these possessed ASA activity judged by the oocyte bioassay. A S A concentrations expressed in this paper reflect the actual concentration of serum albumin (MW 67,000) used. Unless otherwise stated, an albumin isoform which was inactive in the oocyte assay served as the control protein for experiments presented in this paper. Since all vertebrate sera tested so far contained the factor 27. we also used fetal calf serum that is generally used as a tissue culture medium supplement. Blue Agarose chromatography alone was used to deplete fetal calf serum of albumin 28'29. This single step purification yielded a non-bound fraction of serum proteins which were depleted of ASA activity, as judged by the oocyte bioassay. In contrast, the dye-bound ASA fraction produced responses at dilutions over l0 s, and contained most of the ASA activity present in serum. Both the unbound and bound fractions were concentrated to their original volumes using YM-10 membranes in an Amicon concentration cell, and were dialysed against an excess of phosphate buffered physiological saline, pH 7.4.

Morphometric analysis Cultures were routinely photographed at 200 × magnification. Individual frames were further magnified 10x onto a Houston Hipad digitizer pad. To evaluate the NGF-induced early changes, cultures of undifferentiated cells were observed at 400 x magnification with Hoffman modulation contrast microscopy, and flattened (responsive) versus round (non-responsive) shaped cells were counted. In differentiated cells, neurite length, soma area, soma perimeter and number of neurites per cell were measured and analyzed using the Sigma Scan software package (Jandel Scientific, Sausalito, CA) run on an IBM AT computer. Measurements were made on fields containing at least 30 cells, and Student's t test was used to determine whether the various treatments produced statistically significant differences in the various parameters.

Treatment of ASA with antibodie.s to HSA Purified polyclonal monospecific IgG antibodies to HSA made in rabbits and control rabbit antibodies to mouse immunoglobulins (G, A, and M) were purchased from Axell Inc., San Diego, CA, and Zymed, San Francisco, CA, respectively. ASA was incubated with various concentrations of antibodies for 2 h at ambient temperature, and the mixture was added to the culture medium of NGF-differentiated PC12 cells.

Determination of lnsl,4,5P3 in PCI2 cells" Cells were grown in spinner cultures to ensure population homogeneity prior to plating on 35 mm collagen-coated plastic Petri dishes. Cells were allowed to differentiate for 3 days after NGF challenge before Insl,4,5P3 extractions were carried out. Cultures were exposed to 1.5 I,M ASA or inactive control albumin for various lengths of time before incubation was terminated by the addition of 10% perchloric acid to 1.65% final concentration. Insl,4, 5P3 was extracted according to Sharpes and McCar123, and extracts were stored at -80°C until assayed. Three different assay procedures were used to determine Insl,4,5P3 concentrations, all of which employed Insl,4,5P3 binding protein: (t) the commercially available radioligand binding assay kit TRK.1000 (Amersham, Arlington Heights, IL); (2) the commercially available radioligand binding assay NEK-064, (New England Nuclear Corporation, Wilmington, ME); and (3) due to the high cost of these kits, we prepared the Insl,4,5P3 binding protein ourselves from rat cerebellum 4, [3H]Ins1,4,5P3 was purchased from Amersham (TRK.999: spec. act. 22 mCi/mmol). All three binding assays yielded virtually identical Insl,4,5P3 levels when compared to each other. Protein concentrations were determined by the bicinchoninic acid protein assay kit (Pierce Chemical Co., Rockford, IL). RESULTS

Culture of PCt2 cells PC12 cells were grown in Dulbecco's modified Eagle's medium (DMEM) containing 1% fetal calf serum with 11.1 mM D-glucose. For serum-free differentiation studies, cells were seeded on collagen-coated Petri dishes 2° (Vitrogen 100, Collagen Corporation, Palo Alto, CA) in DMEM-N15 serum-free chemically defined medium as described by Skaper et al. 24 at a density of 4104 cells per ml. In some experiments, poly-e-lysine (molecular weight range 70-150 kDa), laminin, and poly-ornithine- (molecular weight range 25-30 kDa) coated plastic Petri dishes (Costar, Cambridge, MA) were used without noticeable differences in culture.

NGF-induced neuronal differentiation of PCl2 cells 2.5S fl-NGF was applied to PC12 cultures at a concentration of

Inhibition o f NGf'-induced early morphological changes by A S A W i t h i n 30 m i n o f e x p o s u r e their spherical morphology lygonal. Subsequently,

to N G F , P C 1 2 cells l o s e

and become

transiently po-

t h e cells b e g i n to s e n d o u t p r o -

c e s s e s , s o m e o f w h i c h will e v e n t u a l l y b e c o m e n e u r i t e s 15. To e v a l u a t e t h e e f f e c t o f A S A o n t h e N G F - i n d u c e d ferentiation,

cells w e r e e x p o s e d

a l o n e , o r to N G F

either to 2 nM

difNGF

p l u s 4,5 ~ M A S A o r i n a c t i v e c o n t r o l

p r o t e i n . A f t e r 30 m i n o f e x p o s u r e to N G F a p p r o x i m a t e l y

295 44% of cells had acquired a flattened, polygonal shape (Fig. 1). This percentage increased to about 89% by 60 min. Cells exposed to N G F in the presence of the inactive control protein had a similar response. In contrast, after 30 min of exposure to N G F in the presence of A S A , 31% of the cells displayed a flattened morphology, whereas by 60 min, only 54% showed a typical N G F - r e sponsive shape.

Retraction of neurites caused by A S A PC12 cells, differentiated in the presence of N G F for 2 days, when exposed to 1.5/~M A S A began to retract neurites within 20 min, and by 60 min the majority of neurites had withdrawn (Fig. 2). In contrast, cultures exposed to the same concentration of inactive control albumin (see Materials and Methods) continued to extend neurites at a rate of about 10 /~m/day, indicating that A S A - i n d u c e d neurite retraction was not simply due to non-specific effects. Substantial neurite withdrawal was observed at doses as low as 750 nM, and retraction was maximal within 2 h. The first noticeable changes after the application of A S A were a decrease in growth cone motility and a shift from a phase dark to a phase bright appearance. The withdrawal of processes was acc o m p a n i e d by a transient swelling which led to blebbing along the retracting neurite (Fig. 2B). Supravital staining of cultures with Trypan blue revealed that cell viability was unaltered (greater than 98%) after 12 h of exposure to A S A . A l t h o u g h the A S A - i n d u c e d neurite withdrawal was usually c o m p l e t e for most cells in a culture, a few cells had neurites which did not retract. Some cells ( - 5 - 7 % ) exposed to A S A d e t a c h e d from

the collagen substrate. To test the possibility that the retraction was due to an interference with substrate adhesion, we examined the effects of A S A on cells which were differentiated on poly-ornithine, poly-L-lysine-, or poly-L-lysine plus laminin-coated culture dishes. A S A was found to cause neurite withdrawal and cell detachment with a similar time-course, regardless of the substrate used. M o r e o v e r , when A S A was left in the culture m e d i u m , d e t a c h e d cells r e - a d h e r e d within 3-5 h and started to re-extend processes. Cells which r e m a i n e d attached also began re-extending neurites by this time. Statistical analysis of neurite lengths, m e a s u r e d from the time-lapse photographs, showed differences in the susceptibility of various neurite populations to A S A . Neurites were classified as either short (1-15/~m) or long ( > 15/~m), and the n u m b e r of processes p e r cell in each group was counted over the course of a 2 h exposure to 1.5 ~ M A S A . Neurites in both length groups showed a rapid shortening, however, the shortest group of processes started to re-grow one hour after the exposure to A S A (Fig. 3A). Cultures treated with the inactive con-

©

O

¢$

100 l

cot +~'u 801 =¢n 8

60

~

40

8 r~ f-

20

30

Time (Minutes)

60

Fig. 1. ASA inhibits the early morphological signs of NGF-induced differentiation. Cells were exposed to either 2 nM NGF alone (gray bars), or to NGF with 4.5 ktM inactive control albumin (open bars), or with 4.5/~M of ASA (filled bars). Cultures were photographed at different times, and cell morphology was assessed. Data represent the mean cell population that was polygonal in shape as a percentage of the total cell population observed for each sample + S.E.M. for 3 separate experiments. *Significantly different from both the NGF and NGF+ inactive albumin groups (P _< 0.05 obtained from a one-tailed Student's t-test).

Fig. 2. ASA-induced retraction of neurites in differentiated PC12 cells. PC12 cells were differentiated as described in methods and exposed to 1.5/~M ASA. A: appearance of a cell group before exposure to ASA. B-D: the same cell group 20, 40 and 60 min, respectively, after adding ASA.

296 A

TABLE I 2o

_T_

Morphometric characteristics of PC12 cells pre- and post-exposure to 1.5 uM ASA, and 2 days after removal of ASA-containing medium

I

III CO +1

Values represent mean + S.E.M.

I

. IL

1.o,

g Z

0 0

0

15

60

Parameter

120

Time (minutes)

B

Pre-exposure

Neurite length (Hm) 22.5+_2.3 Cell area (Hm2) 11.0+_1.2"* Soma perimeter (urn) 31.3 +_ 2.2**

2 Hours" post-exposure

2 Days after removal of active serum albumin

9.2+_1.9",** 22.9 _+ 4.0 9.6 +_ 1.1"*

16.0 +_ 1.2

30.8 +_ 2.1"*

39.9 + 1.8

*P -< 0.05 in comparison with pre-exposure value; **P - 0.05 in comparison with 2 day regeneration value; values of P were obtained using a one-tailed Student's t-test. _T__

UJ CO +1

g U

g

Z

00 15

60

120

Time (minutes)

Fig. 3. Effect of ASA on neurites of different length and culture age. Three-day differentiated cells (A) show a greater susceptibility to ASA (1.5 HM) compared to cells differentiated for 5 days in NGF (B). Neurites are grouped according to their length: 0-15 Hm, open bars; >15 Hm, gray bars. Data points represent means obtained from samples of at least 150 cells.

trol protein showed no statistically significant change in their initial neurite length and population characteristics. It has been shown that serum is less effective in causing neurite retraction in 'older' (5-7 days) than in 'younger' (2-3 days) cultures 25. Accordingly, Fig. 3B shows that the vulnerability of PC12 neurites to A S A also decreased in older cultures of PC12 cells. The most prominent retraction was seen in cultures differentiated for 2 days, while 5-day-old cultures showed only a small, transient decrease in the n u m b e r of short length neurites. The aging effect not only involved the time-course of the retraction, but was also accompanied by an increase in the n u m b e r of ASA-resistant cells. W h e n the medium of ASA-treated cultures was replaced with fresh DMEM-N1 containing NGF, or even if the ASA-containing medium was not replaced, cells that had withdrawn their neurites showed re-extension. For example, the mean length of processes decreased significantly after 2 h of A S A challenge (Table I). W h e n the medium was replaced and cells were allowed to regrow processes

for two days, the mean length acquired was not significantly different from that before exposure to ASA, While PC12 cells exposed to A S A showed no changes in mean cell perimeter and cell area, these parameters increased significantly (compared to the pre-exposure values) after replacing the medium with fresh NGF-containing DMEM-N1 (Table I). I n h i b i t i o n o f A S A action b y a n t i b o d i e s to s e r u m a l b u m i n

Biochemical analysis indicates that the factor responsible for neurite retraction is a form of serum albumin2s. This was supported by the finding that several batches of potyclonal monospecific antibodies to serum albumin caused a dose-dependent block of the ASA-elicited oscillatory currents in oocytes 26'28. Therefore, antibodies to h u m a n albumin were tested for their ability to block the ASA-induced retraction of neurites. ASA purified

TABLE II Blockade of ASA-induced neurite retraction by antibodies, to human serum albumin

3 #M ASA was used in all cases. Values represent mean + S.E.M. Anti-albumin antibodies (pg/ml)

% of original neurite length after 30 min of exposure

200 i00 50 25 12.5 0

99.5 92.8 103.1 70.0 54.2 59.0

_+ 11.1 + 10.0 +- 11.3 + 14.8' + 14.2" _+ 12.8"

*P --< 0.01: values of P were obtained from a one-tailed Student's t-test.

297 12'

A UJ

O~ +I

+l

! i

11111

10'

u

8'

i

6'

II

0

Ouratlon of eXp~lurl (mlnutN)

0

A +l c=

:o: p

+

60

+so

Durdon

lOO

o.

0

0

12.5

25

50

100

of e x p o s u r e ( m l n u t H )

Fig. 5. Effect of bound and unbound Blue Agarose fractions on PC12 cells. Differentiated PC12 cells were exposed for 60 min to 200 ~g/ml of either Blue Agarose 'bound' (filled bars) or 'unbound' serum fractions (open bars), prepared as described in methods. Cultures were photographed before and after incubation, and the average neurite length for cells in each condition was determined. Bars represent mean neurite length + S.E.M., *P