Kidney International, Vol. 58 (2000), pp. 1500–1510

Microsphere-adenoviral complexes target and transducethe glomerulus in vivo

N. STANLEY NAHMAN, JR., THOMAS J. SFERRA, JULIE KRONENBERGER, KATHLEEN E. URBAN,ANNE E. TROIKE, AMY JOHNSON, BETHANY J. HOLYCROSS, GERARD J. NUOVO,and DANIEL D. SEDMAK

Department of Internal Medicine; Department of Molecular Medicine, Children’s Hospital Medical Center; Department ofMedical Biochemistry and Department of Pathology, The Ohio State University, Columbus, Ohio, USA

Microsphere-adenoviral complexes target and transduce the therapy. The development of useful approaches to thera-glomerulus in vivo. peutic gene transfer to the glomerulus depends on at

Background. Developing new treatments for glomerulone- least two essential components: the accurate deliveryphritis makes the glomerulus a logical target for gene therapy.

of DNA constructs to the glomerular capillary and, onMicrospheres may lodge in the glomerulus, and replication-arrival, effective intracellular transfer of the nucleotidedeficient recombinant adenoviruses are potent mediators of

gene transfer. We postulated that adenoviral-microsphere com- sequences of interest. Previous studies have demon-plexes could result in DNA transfer (transduction) into glomer- strated glomerular gene transfer using continuous kidneyular cells in vivo. perfusion with adenovirus [3], renal artery injection of

Methods. Two adenoviruses, each one containing a luciferase liposome complexes [4–7], or arterial injection of trans-or b-galactosidase (b-gal) transgene expression cassette, werefected mesangial cell vectors [8, 9]. Each of these ap-complexed to polystyrene microspheres. To assess in vivo glo-proaches to glomerular gene transfer may be limited bymerular transduction with this tool, male Sprague-Dawley rats

underwent aortic injections with adenovirus linked to 11 or the necessity of invasive surgery or technically complex16 mm diameter microspheres. ancillary procedures. One of the goals of the present

Results. After 48 hours, adenoviral-microsphere complexes study was to develop a minimally invasive and effectiveresulted in transduction of up to 19% of glomeruli per kidney

method for direct glomerular gene transfer.section. Endothelial and mesangial cells were transduced withMicrospheres are recognized to lodge in the glomeru-this approach, and transprotein expression persisted for 21

lus [10] and are thus potential vehicles for shuttling DNAdays. Transduction efficiency was greater in the 16 mm group.For all rats, there was a strong correlation between kidney to the glomerular capillary. Replication-deficient recom-luciferase levels and the number of b-gal-positive glomeruli binant adenoviruses are extremely effective mediators(r 5 0.87), indicating that transgene expression was primarily of gene transfer [11] and would be expected to functionglomerular in location. This was supported by reverse tran-

efficiently if contact with appropriate target tissue canscriptase in situ polymerase chain reaction, which demon-be attained. In the present studies, we postulated thatstrated glomerular localization of the b-gal transgene.

Conclusions. The aortic injection of adenoviral-microsphere combining the tissue-targeting strength of microspherescomplexes transduces the glomerulus in vivo and may be a with the powerful transduction properties of adenovirususeful tool in developing approaches to gene therapy of glomer- would result in DNA transfer into cells of the glomerulusular disease.

in vivo. To address this question, we constructed adeno-viral-microsphere complexes and, via aortic injection,attempted to transduce glomeruli of normal rats. The

Resident glomerular cells may mediate evolution of results of these studies indicate that an adenoviral-many forms of acute and chronic glomerulonephritis microsphere shuttle system is well tolerated and results[1, 2], making the glomerulus a logical target for gene in efficient transduction of glomerular endothelial and

mesangial cells. This method may prove to be a useful newapproach for targeting gene therapy to the glomerulus.Key words: gene therapy, acute glomerulonephritis, chronic glomerular

disease, nucleotide sequences, mesangial cell vectors, glomerular genetransfer, replication-deficient recombinant adenovirus.

METHODSReceived for publication September 4, 1998 Adenoviral constructsand in revised form April 4, 2000Accepted for publication April 25, 2000 Preparation of replication-deficient recombinant ade-

novirus. Replication-deficient recombinant adenovirus 2000 by the International Society of Nephrology

1500

Nahman et al: Microsphere-adenoviral complexes and transduction 1501

used in these studies was prepared as we have described Co. (Franklin Lakes, NJ, USA) using RPMI 1640 fromGIBCO BRL (Gaithersburg, MD, USA) supplementedin detail elsewhere [12]. In brief, viral stocks of adenovi-

rus-containing expression cassettes composed of the cy- with 10% FCS, penicillin (100 m/mL), streptomycin(100 m/mL), and amphotericin-B (0.25 mg/mL; GIBCO).tomegalovirus immediate early promoter fused to the

lacZ reporter gene (rAvb-gal; a generous gift of Dr. Phil Approximately 105 cells were transduced with 105 adeno-viral-microsphere complexes dissolved in the previouslyJohnson, Children’s Hospital, Columbus, OH, USA), or

the distal 459 base pairs of the human transforming mentioned media supplemented with 2% FCS, and/orthe washes derived from the preparation of the adenovi-growth factor (TGF) gene promoter (a generous gift of

Dr. Seong Kim, National Institutes of Health, Bethesda, ral-microsphere complexes. After two hours, the tissueculture supernatant was removed and replaced with fullMD, USA) fused to a luciferase reporter gene (rAvTGF)

were used in these studies [13]. media. Twenty-four hours later, soluble transgene activ-ity (luciferase) from cell extracts was determined. InThe rAvb-gal reporter gene codes for the enzyme

b-galactosidase (b-gal), and the rAvTGF construct gen- some studies, tissue transgene expression was assessedby staining fixed cells for b-gal activity.erates the firefly luciferase transprotein. The adenoviral

amplification and purification procedure is summarized In vivo transduction using adenoviral-microsphere com-plexes. Male Sprague-Dawley rats weighing 300 to 350 gas follows: Viral recombinants were amplified by mixing

a sample from stock adenovirus with supplemented (Harlan-Sprague Dawley, Indianapolis, IN, USA) wereanesthetized with intraperitoneal pentobarbital (50 mg/kgEarle’s Minimum Essential Medium (EMEM) from

Quality Biologicals (Gaithersburg, MD, USA). The solu- body weight). The left carotid artery was identified viaa small anterior, cervical-thoracic incision and was can-tion was applied to confluent 293 cells obtained from

ATCC (Rockville, MD, USA). After evidence of exten- nulated using polyethylene tubing (PE-50), which wasadvanced approximately 3 to 5 cm into the aorta, fol-sive cytopathic effect (3 to 4 days), the virus was recov-

ered following three freeze-thaw cycles and centrifuga- lowed by injection of 1 mL of PBS containing the adeno-viral-microsphere complexes. The proximal carotid ar-tion. The viral recombinants were purified with cesium

chloride ultracentrifugation and dialyzed [14]. The puri- tery was ligated as the catheter was removed, and theincision was closed with wound clips. Animals were sacri-fied virus was resuspended in 10% sterile glycerol and

stored at 2708C. ficed after 48 hours or 7, 14, or 28 days, and samples ofkidney, spleen, liver, and lung were taken for assessmentTo determine the titer (TCID50) of the purified virus,

sample dilutions (1023, 1026, and 1029 through 10212) in of transgene activity.EMEM 1 2% fetal calf serum (FCS) and l-glutamine

Determination of transgene expressionwere made and used to inoculate a 12-well plate con-taining confluent 293 cells. After three hours, the media Luciferase activity from cultured cells and transduced

kidneys. Luciferase activity from transduced mesangialwere changed, and each well was examined daily forevidence of cytopathic effect. Appropriate transgene ex- cells was determined from cell extracts obtained as we

have described [12]. In brief, wells were washed with PBS,pression from transduced cells was assessed in parallelwells [12]. followed by the addition of a single-step cell lysis buffer

(Promega, Madison, WI, USA), scraped with a rubberAdenoviral-microsphere complexes. Adenoviral-micro-sphere complexes were prepared by coincubating rAvb- policeman, and exposed to a single freeze-thaw cycle,

and the samples were frozen at 2708C until assayed.gal and rAvTGF (109 particles/mL for each construct)with 106 11 or 16 mm diameter fluorescent polystyrene Tissue luciferase levels from kidney, spleen, liver, and

lung of transduced rats were determined after snap freez-microspheres from Duke Scientific Corporation (PaloAlto, CA, USA) in borate buffer, pH 8.5. After two ing the sample in liquid nitrogen, pulverizing at 2708C,

and solubilizing in the presence of cell lysis buffer (2.5hours at 48C, the adenoviral-microsphere complexeswere washed four times in phosphate-buffered saline mL/g tissue). After incubation for 30 minutes at 48C and

centrifugation at 1500 r.p.m. for 10 minutes at 48C, the(PBS; pH 7.20) and maintained at 48C on a rotatingshaker until utilized for transduction. supernatant was recovered and stored at 2708C. Lucifer-

ase levels were determined by mixing 20 mL of cell orTransduction protocols tissue extract with 100 mL of luciferase assay substrate

from Promega and reading immediately in a Lumat LBIn vitro transduction of human mesangial cells withadenoviral-microsphere complexes. Mesangial cells were 9501 luminometer (Wallac, Gaithersburg, MD, USA).

Kidney expression of the b-gal transgene in transducedisolated from adult human kidneys, deemed unaccept-able for transplantation, using sieving techniques and rats. Kidney expression of the b-gal transgene was docu-

mented in vivo using reverse transcription-polymerasecollagenase digestion, as we have previously described[15, 16]. The isolated mesangial cells were cultured in chain reaction (RT-PCR) and RT in situ PCR.

(a) RT-PCR. The total RNA from a single kidney6- or 12-well plates from Falcon, Becton Dickinson and

Nahman et al: Microsphere-adenoviral complexes and transduction1502

Fig. 1. In vitro transduction of cultured human mesangial cells. (A) Soluble luciferase activity from mesangial cells transduced with microspheres(MS) lacking adenovirus, microspheres complexed to 2 adenoviruses, 1 each bearing a luciferase or lacZ expression cassette (MS 1 rAv), or consecutivewash solutions from MS 1 rAv immediately following the adsorption reaction. Luciferase activity from cells transduced with MS 1 rAv (after 4washes) was more than twofold over background (MS only). Luciferase activity from cells transfected with the third wash solution was no differentfrom background, suggesting that at this point, all transduction was mediated by adenovirus complexed to microspheres and not from free virus insolution (*P , 0.05 when compared with wash 1). (B) Histologic appearance of b-gal transgene activity (blue color) in transduced mesangial cells.Most transduced cells are associated with a microsphere (arrows). The presence of wild-type (“untransduced”) cells (purple color) suggests that themajority of transduction was associated with virus bound to a microsphere since free virus will avidly transduce wild-type mesangial cells [12] (325).

Fig. 3. Histologic appearance of b-gal trans-gene protein expression (blue color) from theglomeruli of transduced rats. Microspheres areindicated by the black arrows. (A–E ) Repre-sentative glomeruli exhibiting a spectrum oftransduction patterns. These patterns and theimage analysis characteristics of each glomer-ulus are indicated in Table 2. (F) Higher powerimage of the glomerulus illustrated in (E) indi-cating cytoplasmic staining for the b-gal trans-gene protein, with relative sparing of the nu-clei (white arrows).

from either a wild-type or transduced rat, or transduced of 0.1 mg of oligo(dT)18 in a total volume of 13 mL. Afterquick chilling in an ice bath, 6 mL of a mixture containing293 cells (as a positive control), was extracted using the

method of Chomczynski and Sacchi [17]. Single-stranded 70 mmol/L Tris, 160 mmol/L KCl, 8 mmol/L MgCl2,3 mg/mL bovine serum albumin (BSA), 0.33 mmol/L di-cDNA was synthesized from total RNA using an oli-

go(dT)18 primer and recombinant Moloney murine leu- thiothreitol (DTT), 2 mmol/L each of dNTPs, and 33m/mL RT were added. After 10 minutes at room tempera-kemia virus RT from GIBCO. One microgram of total

RNA was heated to 708C for 10 minutes in the presence ture, 428C for 50 minutes, 908C for 5 minutes, and place-

Nahman et al: Microsphere-adenoviral complexes and transduction 1503

ment on ice, 2 m of RNase H were added, and the mixture utes at 48C. The sections were washed twice in PBSwas incubated at 378C for 20 minutes. PCR amplification and stained in a 100 mmol/L sodium phosphate solutionof target cDNA was as follows: 1 mg of cDNA and 100 containing 1.3 mmol/L MgCl2, 3 mmol/L K3Fe(CN)6,nmol/L of primers recognizing the 59 forward terminus 3 mmol/L K4Fe(CN)6, and 1 mg/mL 5-bromo-4-chloro-(59-CCGTCGTTTTACAACGTCGTGA-39) and 39 re- 3-indolyl-b-d-galactoside (X-gal) at room temperatureverse terminus (59-ATGTGAGCGAGTAACAACCG for 14 to 16 hours. The sections were counterstainedTCGGATTCT-39) of b-gal mRNA (GIBCO) were with nuclear fast red. The sections were examined bymixed and cycled at 948C 3 1 minute, 558C 3 1 minute, visual inspection and analyzed using digital analysis.728C 3 1 minute for a total of 30 cycles, followed by a Light and electron microscopic analysis of transducedfinal extension at 728C for 10 minutes. Equal volumes tissue. Each kidney was examined under light micros-of each PCR product were evaluated with 1.5% agarose copy with the observer blinded to the tissue type. Ingel containing 0.5% ethidium bromide. The product rep- this analysis, all kidneys were examined for evidenceresenting the b-gal transgene was identified at 355 bp. of ischemia and the anatomic (glomerular, tubular, or

(b) RT in situ PCR. Four micron cryostat kidney sec- peritubular) distribution of b-gal transgene expression.tions from a wild-type and transduced rat were fixed in In b-gal-positive glomeruli, the cellular expression of the10% buffered formalin for two days and washed in b-gal transgene was determined by examination of 1 mRNase-free water and 100% ethanol. After air drying, thick sections stained for four hours using X-gal.the tissues were protease digested (the optimal digestion Glomerular transduction was defined as the presencetime was determined using nonspecific incorporation of of b-gal activity inside Bowman’s capsule. If b-gal-posi-10 mmol/L digoxigenin dUTP) [18, 19], incubated over- tive glomeruli were noted, the total glomerular countnight in 10 U RNase free-DNase (Boehringer Mann- and number of transduced glomeruli were determined.heim, Indianapolis, IN, USA), and subjected to one step To determine the intraglomerular distribution of trans-RT-PCR using the rTth system and digoxigenin dUTP duction in each b-gal (1) glomerulus, the location (inner,as we have described [18, 19]. Using the previously men- middle, or outer third of the glomerular tuft circumfer-tioned primer sequences for b-gal, 20 cycles of PCR were ence) of b-gal transgene staining was recorded.performed, and the slides washed at high stringency (608C The extent of b-gal transgene expression from kidneysfor 10 min in 15 mmol/L NaCl containing 2% bovine transduced with adenoviral-microsphere complexes wasserum albumin) and the digoxigenin-labeled, target-spe- quantitated using digital image analysis. A 24-bit, 325cific cDNA detected after a 30 minute incubation at 378C magnified digital image of every b-gal-positive glomeru-with 1:200 antidigoxigenin-alkaline phosphotase conju- lus from each kidney was prepared using a SPOT digitalgate (Boehringer Mannheim) followed by exposure to

camera (Diagnostic Instruments, Sterling Heights, MI,nitroblue tetrazolium and bromochloroindolyl phosphate

USA). The surface area of b-gal-positive regions in eachchromogens (Enzo Biochemicals, Farmingdale, NY, USA).glomerulus was determined with an image analysis soft-The tissues were subsequently counterstained with nu-ware program from Jandel (SigmaScanPro, San Rafael,clear fast red. Controls for RT in situ PCR included noCA, USA). In addition, the total surface area of theDNase treatment, the omission of primers, the use ofglomerulus under study was determined, allowing calcu-irrelevant human papillomavirus-specific primers, andlation of the percentage of the total glomerular surfaceRNase digestion prior to RT in situ PCR. b-gal mRNAarea transduced. Electron microscopic analysis of kid-was indicated by the presence of a blue precipitate.neys transduced with 16 m diameter beads was performedHistologic preparations for b-gal activity. Histologicusing methods similar to what we have described pre-preparations of transduced mesangial cell cultures orviously [12]. In brief, tissue was obtained 24 hours follow-organ samples from transduced rats were prepared asing transduction. After bisecting the left kidney in thewe have described [12]. In brief, mesangial cells werelongitudinal plane, one half of the tissue mass was dis-grown in chamber slides (Nalge Nunc International, Na-sected into 1 mm cubes, fixed in glutaraldehyde, imbed-perville, IL, USA), and after washing with PBS, theyded in Epon, sectioned at 50 to 60 nm, and stained withwere fixed in 5% formalin and 0.05% glutaraldehyde aturanylacetate and lead acetate. The remaining portionroom temperature for five minutes, washed, stained withof the kidney was sectioned and stained for b-gal activity.an X-gal solution from GIBCO, and examined by visual

inspection. Kidneys from transduced rats were removedStatistical analysisat sacrifice and bisected longitudinally, and a portion of

All data are expressed as mean 6 SEM. Groups werethe organ was frozen in cryoprotective medium (OCT;compared using Student’s t-test, and differences wereSakura Finetek, Torrance, CA, USA) and stored atconsidered significant at P value , 0.05. Linear regres-2708C until sectioned. Cryosections (1 to 6 mm) weresion analysis was used to compare luciferase levels andfixed with 2% formaldehyde and 0.2% glutaraldehyde

in a 100 mmol/L phosphate buffer (pH 7.4) for five min- histologic expression of the b-gal transprotein.

Nahman et al: Microsphere-adenoviral complexes and transduction1504

Fig. 5. Cellular localization of the b-gal trans-protein (blue color) from transduced rats. (A)Transduced endothelium (arrow) in directproximity to a microsphere (MS). (B ) Trans-duced mesangial cells (arrow) in a segmentalportion of a glomerular tuft (31250).

Fig. 7. Expression of the b-gal transgene from transduced rat kidney using RT in situ PCR. (A) Low-power view of a representative kidneysection from a rat transduced with microspheres complexed to recombinant adenovirus containing a CMV/lacZ expression cassette. b-gal mRNAis indicated by the blue pigment, which is localized to the glomerulus (large arrow). There is no signal in tubular cells remote to the transducedglomerulus (small arrow; 3400). (B ) High-power image of (A) indicating numerous b-gal-positive cells (large arrow). Cells negative for b-galactivity are indicated by the small arrow (31000). (C ) High-power image of a tubule from (A) showing no b-gal activity (31000). (D ) Representativeglomerulus from a rat injected with microspheres only. There is no evidence of b-gal activity (3400).

Nahman et al: Microsphere-adenoviral complexes and transduction 1505

RESULTS

In vitro transduction of cultured humanmesangial cells

Microsphere-adenoviral complexes efficiently trans-duced human mesangial cells in culture. Figure 1A dem-onstrates luciferase transgene activity from mesangialcells transduced with microspheres only or microsphere-adenoviral complexes. Microspheres containing the ade-novirus demonstrated over a twofold increase in lucifer-ase activity when compared with microspheres alone(P , 0.05). To confirm that transduction resulted fromadenovirus complexed to microspheres and not from free(unbound) virus, microsphere-adenoviral complexes wereprepared in the borate adsorption buffer, and the super-natant from consecutive washes was used to transducemesangial cells. The luciferase activity generated withthis approach is a reflection of free virus present in thesupernatant after each wash. Luciferase levels progres-sively declined with each wash, with transgene activityfalling below that of complexed microspheres by thethird wash.

The histologic appearance of cultured mesangial cellstransduced with microsphere-adenoviral complexes is in-dicated in Figure 1B. The b-gal staining pattern is primar-ily cytoplasmic, as we have previously shown with freeadenovirus [12]. Virus bound to the microspheres likelymediated transduction, because luciferase activity fromcells transduced with the adenoviral-microsphere com-plexes was significantly greater than levels from cellsincubated the final wash solution. This implies that themajority of the virus was bound to the microspheres andnot free in solution. We have previously shown that asfew as one viral particle per cell can result in transductionand detectable transprotein levels [12].

In vivo transduction of rat kidney

Table 1 presents clinical data and transgene activity forindividual rats transduced with 16 or 11 mm microspherescomplexed to adenovirus. All rats tolerated the procedurewithout evidence of postoperative morbidity or mortality.

At the time of transduction, the weight of rats trans-duced with 11 mm microspheres was slightly but signifi-cantly greater than rats transduced with 16 mm microspherecomplexes. Renal function was assessed by measuringserum creatinine levels. There were no differences in base-line serum creatinine concentrations between the twogroups; however, 48 hours following transduction, bothgroups had modest but significant increases in serumcreatinine.

On histologic evaluation, all kidneys were normal inappearance and were without evidence of ischemia. Therewere no differences between the groups in total numberof glomeruli per kidney section; however, there was trend

Tab

le1.

Clin

ical

data

and

tran

sgen

eac

tivi

ties

from

indi

vidu

alra

tstr

ansd

uced

wit

had

enov

iral

-mic

rosp

here

com

plex

es

Seru

mcr

eati

nine

leve

lbH

isto

logi

cas

sess

men

tof

bga

lex

pres

sion

Imag

ean

alys

isof

Rat

num

ber

MSa

TD

Tot

alb

-gal

(1)

MS

(1)

glom

erul

ib

-gal

(1)

surf

ace

LC

Fby

grou

pin

ject

ed(N

)w

eigh

tcP

re-T

DP

ost-

TD

glom

erul

igl

omer

uli

%%

b-g

al(1

)dar

ea%

ele

velf

16m

mm

icro

sphe

res

100

53

105

342

0.4

0.5

181

5(3

)19

(26)

7.7

61

(36

1)37

110

110

634

80.

40.

618

929

(15)

54(5

4)34

65

(13

62)

905

102

106

345

0.4

0.5

185

5(3

)14

(36)

176

4(6

62)

476

103

23

106

340

0.4

0.6

160

31(1

9)64

(48)

256

4(1

16

2)68

810

42

310

634

60.

40.

615

029

(19)

84(3

5)24

64

(12

61)

650

Mea

n6

SEM

344

61g

0.4

60

0.6

60.

02h

173

68

206

6(1

26

4)47

613

(40

65)

g22

64

(96

2)g

618

692

11m

mm

icro

sphe

res

105

53

105

351

0.5

0.5

166

2(1

)21

(10)

4.6

63

(1.3

61)

332

107

106

354

0.4

0.6

184

1(1

)4

(25)

4.2

60

(1.5

60)

240

108

23

106

346

0.5

0.6

208

31(1

5)10

5(3

0)17

64

(76

1)11

4510

92

310

635

60.

40.

513

99

(6)

65(1

4)6.

26

1(3

60)

433

Mea

n6

SEM

352

62

0.45

60.

030.

556

0.03

h17

46

1511

67

(66

3)49

623

(19

65)

86

3(3

61)

538

620

6aM

icro

sphe

res

bE

xpre

ssed

inm

g/dL

;pre

-TD

and

post

-TD

refe

rto

leve

lsin

seru

mob

tain

edpr

ior

to,o

r48

hour

saf

ter

tran

sduc

tion

,res

pect

ivel

ycW

eigh

tin

gram

sat

the

tim

eof

tran

sduc

tion

(TD

)dN

umbe

rof

glom

erul

ico

ntai

ning

am

icro

sphe

re;t

henu

mbe

rin

pare

nthe

ses

repr

esen

tsth

epe

rcen

tof

MS

posi

tive

glom

erul

iex

pres

sing

bga

lac

tivi

tyeM

ean

6SE

Mof

b-g

alpo

siti

vesu

rfac

ear

ea(3

104

pixe

lssq

uare

d)fo

ral

ltr

ansd

uced

glom

erul

ifr

oma

give

nki

dney

.T

heva

lue

inpa

rent

hese

sre

pres

ents

the

perc

ent

ofth

egl

omer

ular

surf

ace

area

that

stai

ned

posi

tive

for

b-g

alf L

ucif

eras

e(L

CF

)le

vel

inar

bitr

ary

unit

s;ea

chva

lue

was

norm

aliz

edto

the

wei

ght

ofth

eki

dney

tiss

uem

ass

assa

yed

gP

,0.

05w

hen

com

pare

dto

11m

mtr

eatm

ent

grou

phP

,0.

05w

hen

com

pare

dto

the

pre-

TD

leve

l

toward a greater number of b-gal-positive glomeruli in

Nahman et al: Microsphere-adenoviral complexes and transduction1506

Table 2. Transduction and image analysis patterns of the glomerulishown in Figure 3

Figure 3 Microsphere Location of b-gal positivepanel diameter lm b-gal staininga surface areab

A 11 Middle 4.2 (1.5)B 16 Outer 18 (6.9)C 16 Inner 39 (8.7)D 16 Outer 71 (28.1)E 16 Middle 55 (23)

a Refers to the inner, middle or outer one third of the glomerular tuft circumfer-ence

b 3103 square pixels. The number in parentheses refers to the percent of thetotal glomerular tuft surface area expressing b-gal activity. A description of theimage analysis technique is in the text.

in several glomeruli and are usually contiguous with anarea of positive staining for b-gal. We postulate thatFig. 2. Correlation between soluble kidney luciferase levels and histo-transduction occurs from adenoviral contact with targetlogic appearance of the b-gal transgene for individual rats transduced

with adenoviral microsphere complexes. A positive correlation (r 5 cells, which is facilitated by the embolized microsphere.0.87, P , 0.05) between luciferase level and the number of glomeruli Examples of transduction of each third of the glomeru-exhibiting b-gal transgene activity [b-gal (1) glomeruli] was demon-

lar tuft circumference are represented in Figure 3; how-strated.

ever, the majority of the transduction events occurredin the outer one third of the tuft. The distribution ofb-gal staining for all transduced glomeruli in this study

the 16 mm group (20 6 6 vs. 11 6 7 for 16 mm and indicated that 11 mm microspheres almost exclusively11 mm groups, respectively). On average, both groups localized to the outer third of the glomerular tuft (89 6exhibited 47 to 49 microsphere-positive glomeruli per 15%). In contrast, 28 6 17% and 69 6 16% of the 16 mmkidney section, but the percentage of these glomeruli microspheres were localized to the middle and outerexhibiting b-gal activity was significantly greater in the thirds of the tuft, respectively.16 mm group (40 6 5 vs. 19 6 5% for 16 mm vs. 11 mm, The diffuse nature of the b-gal staining in Figure 3respectively). In the present studies, the expression of suggested that all cell types making contact with a com-b-gal was limited to the glomeruli and was not demon-

plexed microsphere may be susceptible to transduction.strable in either the tubules or the interstitium. In con-

Thus, microspheres embolized in the glomerular capil-trast, the injection of free virus into the renal artery

lary would be expected to make contact with both endo-results primarily in vascular transduction with sparing ofthelial and mesangial cells. To confirm the location ofthe glomerulus [20, 21].an embolized microsphere relative to target cells, we per-When image analysis was used to compare glomerularformed electron microscopy on a kidney from a rat sub-surface area of b-gal expression between the two groups,jected microsphere-adenoviral transduction according tothere was a threefold increase in surface area stainingthe previously mentioned transduction protocol. The re-and the percentage of glomerular area expressing the b-sults of this experiment are shown in Figure 4, which indi-gal transgene (9 6 2 vs. 3 6 1% for 16 mm vs. 11 mmcates the proximity of both endothelial and mesangialgroups, respectively). There were no differences in stain-cells to the microsphere. Transduction of both glomeru-ing intensity, but there was a significant increase in b-lar endothelial and mesangial cells is shown in Figure 5.gal-positive staining area (22 6 4 vs. 8 6 3 3 104 pixels

To demonstrate that the b-gal transgene was success-squared) for 16 versus 11 mm treated rats, respectively.fully delivered to the kidney, RT-PCR and RT in situ PCRThere were no differences in tissue luciferase levelsfor the lacZ gene sequence was performed on kidneysbetween rats injected with 16 or 11 mm microspheres;isolated from rats injected with microsphere-adenoviralhowever, there was a trend toward higher levels in thecomplexes or control animals treated with microspheres16 mm group (618 6 92 vs. 538 6 206). There was a strongalone. In rats transduced with microsphere-adenoviralcorrelation between luciferase level and both the numbercomplexes, RT-PCR demonstrated the presence of theof transduced glomeruli and b-gal staining area (Fig. 2).lacZ nucleotide sequences in the kidney (Fig. 6), andThe histologic appearance and b-gal transgene expres-RT in situ PCR indicated glomerular localization of thesion for several representative glomeruli are shown intransgene (Fig. 7). Using both techniques, there was noFigure 3 and the patterns are analyzed in Table 2. Asb-gal expression in control rats. The other controls forcan be seen, the glomeruli are histologically normal and

without evidence of ischemia. Microspheres are evident RT in situ PCR were negative, and RNase digestion

Nahman et al: Microsphere-adenoviral complexes and transduction 1507

Fig. 4. Electron micrograph of the glomeru-lus from a rat transduced with 16 mm micro-spheres (MS) complexed to adenovirus. (A)Microsphere lodged in glomerular capillary.The proximity of mesangial cells (MC) to themicrosphere is evident (31800). (B ) Higherpower view of the same field showing contactbetween microsphere and the endothelium. Theepithelial foot processes (closed arrow) and uri-nary space (asterisk) are indicated (34500).

prior to RT in situ PCR led to a loss of b-gal signal (datanot shown).

Duration and organ distribution oftransgene expression

The duration of transgene expression (luciferase activ-ity) from transduced kidneys was assessed for up to 28days. As can be seen in Figure 8, rats transduced withmicrosphere-adenoviral complexes exhibited a signifi-cant increase in kidney luciferase levels for up to 21 days.Luciferase activity was no different than backgroundby 28 days. These data are in agreement with previousreports [20].

To assess the distribution of transduction followingaortic injection, luciferase activity was assessed from kid-ney, spleen, liver, and lung. As can be seen in Figure 9,transgene expression was limited to the kidney andspleen, and was no different from background levels inboth liver and lung.

DISCUSSION

This study is, to our knowledge, the first to demon- Fig. 6. Expression of the b-gal transgene from transduced rat kidney.strate that a systemic injection of replication-deficient RT-PCR was used to amplify total RNA from a rat transduced with

unmodified (lacking virus) microspheres (rat MS only) or an animalrecombinant adenovirus can successfully target andinjected with microspheres complexed to adenovirus bearing the b-galtransduce the glomerulus. The method uses adenovirus gene (rat MS-rAvb-gal). The b-gal signal from rat MS-rAvb-gal (arrow)

complexed to polystyrene microspheres that, through is of identical size to that amplified from 293 cells transduced with theb-gal gene [b-gal (1) cont]. There is no signal demonstrable from the ratembolization of the glomerular capillary, facilitate viralinjected with uncoated microspheres. Control lanes include the following:

infection of target cells. The procedure was well tolerated Markers, l HindIII markers; Cont RNA, a control for reverse transcrip-tion [an 891 bp RNA sequence from the chloramphenicol acetyltransfer-and resulted in successful transduction of up to 19% ofase gene (provided by the vendor)]; b actin, a control for the presenceglomeruli.of human mesangial cell RNA; and 100 mers, 100 bp ladder markers.

Other studies have reported successful gene transferinto the glomerulus using continuous organ perfusion withadenovirus [3], liposome complexes [4–7], or transfectedmesangial cell vectors [8, 9]. Continuous renal perfusion duced glomeruli was not investigated. Thus, the relativewith adenovirus resulted in successful b-gal transgene pro- contribution of passive deposition of b-gal synthesized bytein expression in up to 85% of glomeruli [3]. However, transduced liver cells is unknown [22].

Dexoyoligonucleotides encapsulated in liposomal com-in this study, the presence of the b-gal transgene in trans-

Nahman et al: Microsphere-adenoviral complexes and transduction1508

Fig. 8. Duration of kidney transgene expression. Kidney transgene Fig. 9. Organ distribution of transgene expression from transducedexpression (luciferase activity) was measured in rats transduced with rats. For each organ, the mean luciferase level (N 5 4 rats) was calcu-microsphere-adenoviral complexes at each time point beginning 48 lated and compared using analysis of variance. *P , 0.05 when com-hours after transduction (d), (N 5 4 rats/time point). The open circles pared to liver and lung. Beads 5 kidney luciferase level from a rat(s) represent background kidney luciferase levels from rats injected injected with uncoated microspheres.with uncoated microspheres (N 5 3 rats/time point). When comparedwith controls, animals injected with microsphere-adenoviral complexeshad significantly higher luciferase levels for 21 days. There was a gradualloss of luciferase activity over time with no difference in luciferase

tery cannulation. Finally, the present method may allowlevels 28 days following transduction (*P , 0.05).for further characterization of in vivo transductionevents occurring at the adenovirus–target cell interface.For example, fluorescent adenovirus [23] may be com-

plexes resulted in efficient glomerular gene transfer within plexed to microspheres and used to track subcellular20 minutes [5], two hours [6], and four days [4] of the events associated with transduction in vivo.injection. Transgene activity was sufficient to suppress In the present studies, adenovirus complexed to 16 mmglomerular matrix accumulation [5] and transcription microspheres resulted in higher transduction efficiencyfactor activity [6]. This approach may be limited by tran- than virus complexed to 11 mm particles. This was re-sient transgene expression and complexities associated flected by (1) a greater percentage of microsphere-posi-with vector preparation [7]. tive glomeruli exhibiting transduction (40 6 5 vs. 19 6 5

The direct renal artery injection of transfected mesan- for 16 mm vs. 11 mm, respectively), (2) a larger portiongial cells was associated with selective glomerular gene of the glomerular surface area staining positive for thedelivery, high transduction efficiency, and sustained trans- b-gal transprotein, and (3) by trends toward a greatergene expression for up to four weeks [8]. Transformed number of transduced glomeruli and higher tissue lucif-mesangial cells occupied the glomerular capillary, and erase levels. The lack of difference in tissue luciferasesome cells were shown to exhibit migration into the glo- level between the two microsphere groups is due, at leastmerulus. Similar to our studies, a decrement in renal in part, to the small number of animals studied, combinedfunction was noted in association with glomerular embo- with the highly efficient transduction results from ratlization, and concerns about long-term histologic change #108. Differences in luciferase levels are significant if thishave been raised [9]. animal is excluded from the analysis (data not shown).

The information presented in the current studies dem- The differences in transduction efficiency between theonstrates a new approach to glomerular gene transfer two sizes of microspheres may have resulted from (1)and offers several advantages over previous methods. an increased frequency of glomerular embolization byFirst, preparation of the adenoviral-microsphere com- 16 mm microspheres, (2) the smaller size of the animalsplexes is technically simple and highly reproducible. In in the 16 mm group (which was a random event), and/oraddition, the method uses widely available commercial (3) the larger surface area of the 16 mm particles (804 vs.components. Second, the surgical approach is straightfor- 380 mm2). It is unlikely that the increase in transductionward and not technically demanding. Aortic injection of efficiency observed with the 16 mm microspheres resultedthe microsphere-adenoviral complexes takes less than from increased frequency of glomerular embolization,20 minutes and is associated with no operative morbidity because there were no differences between the twoor mortality. In contrast, all of the above-mentioned groups in the number (and percentage) of glomeruli con-

taining a microsphere. Alternatively, the smaller size ofstudies necessitated laparotomy with selective renal ar-

Nahman et al: Microsphere-adenoviral complexes and transduction 1509

the 16 mm-treated rats may have been associated with the glomeruli of our transduced animals was the result ofviral infection of the glomerulus followed by local trans-a decrease in the volume of distribution of injected mi-

crospheres and thus a higher concentration of micro- gene protein production, and not a consequence of hepaticsynthesis, with secondary deposition in the glomeruli [22].spheres delivered to each glomerulus per unit volume

of blood. We postulate that this is unlikely, however, The present studies demonstrated substantial glomer-ular transgene expression within 48 hours of transduc-on the basis that the number of microsphere-positive

glomeruli was no different between the two groups. We tion, sustained transprotein levels for 21 days, and even-tual loss of transgene activity by day 28. These data arespeculate that the most plausible explanation for the

difference in transduction rate between the two groups in agreement with previous studies using the injection offree adenovirus [20]. We postulate that adenoviral-micro-was the increased surface area of the 16 mm particles.

A larger surface area would be expected to bind more sphere complexes composed of adenoviruses exhibitingprolonged transgene expression [25, 26] could be success-virus and thus confer delivery of a larger number of viral

particles per glomerulus. In addition, the larger particles fully used to transduce the glomerulus for therapeuticindications necessitating prolonged transgene activity.would theoretically promote transduction by allowing

more virus to make contact with a greater surface area The decrease in glomerular filtration rate observedfollowing injection of microsphere-adenoviral complexesof target cells. In vitro, the dose of virus and the duration

of contact with target cells are essential variables mediat- likely resulted from (1) glomerular ischemia from mi-crosphere embolization and (2) local inflammatory re-ing effective viral infection [12, 24].

This work demonstrates a strong correlation between sponses induced by recombinant adenoviruses [27]. Thelatter effect will likely be minimized with the use of lesstotal kidney luciferase levels and b-gal-positive glomer-

uli. This finding suggests that nearly all of the transgene immunogenic viruses [25, 28].Glomerular ischemia from microsphere embolizationexpression was glomerular in origin. However, a substan-

tial portion of transgene expression could be extraglo- limits the applicability of polystyrene microspheres as arealistic clinical tool and necessitates the developmentmerular if (1) rAvTGF (producing luciferase) transduced

tubulointerstitial or vascular cells with greater affinity of biodegradable microsphere vehicles. In this regard,biodegradable microspheres composed of albumin, ca-than rAvb-gal and (2) there was a significant amount of

free (not bound to a microsphere) rAvTGF available to sein, or polyesters may be used for in vivo drug delivery[29–31]. In addition, we have shown that microspherestransduce extraglomerular sites. Although unlikely, we

cannot exclude differences in cellular tropism for each composed of the biodegradable polyester poly(lactide-co-glycolide) complex with adenovirus and efficientlyviral construct. We postulate, however, that there was

no unbound rAvTGF, due to the lack of hepatic trans- mediate transduction in vitro (unpublished observa-tions). We hypothesize that the ideal microsphere parti-duction (as seen with the injection of free adenovectors)

[22]. The lack of free virus makes it less likely that selec- cle would (1) be of approximately 16 mm in diameter,(2) occlude the glomerular capillary long enough to allowtive tubulointerstitial or vascular transduction could

have occurred. The glomerular localization of the b-gal complete adenoviral transduction, and (3) exhibit rapiddissipation for restoration of normal glomerular bloodand luciferase transproteins contrasts markedly with re-

nal artery injection of free adenovirus in which the ma- flow. The current studies have shown the feasibility ofa microsphere-based system for glomerular targeting andjority of transgene expression in the kidney is localized

to the peritubular capillaries and tubular cells, with no provide important data for the development of a biode-gradable particle for use in this approach.apparent glomerular transduction [20, 21].

In the present studies, transprotein expression was In these studies, up to 19% of glomeruli were success-fully transduced. At this level of transduction, approxi-limited to the kidney and spleen, with negligible activity

in liver and lung. We postulate that this occurred because mately 12% of the surface area of these glomeruli exhib-ited b-gal transprotein expression. It is unclear which(1) the majority of the adenovirus remained bound to

microspheres and (2) the capillary beds of the spleen of these indicators of transduction is of greater clinicalimportance in the development of gene therapy for glo-and kidney are readily accessible following aortic injec-

tion. In contrast, the lack of transgene activity in liver merular disease. The physical characteristics of the thera-peutic construct are likely critical variables. For example,and lung likely resulted from microsphere-adenoviral

trapping in capillary beds proximal to each organ. if the transprotein is secreted by target cells, transductionof only a few glomerular cells may be adequate to attainIn these studies, we used RT-PCR and RT in situ PCR

to document expression of the b-gal transgene in kidney a therapeutic effect in the entire glomerulus. In contrast,the effect of a therapeutic antisense sequence or ribo-and glomerulus, respectively, from transduced rats. Con-

trol animals injected with microspheres lacking the virus zyme may be limited to only to those cells transduced,thus necessitating much higher levels of transduction.did not exhibit b-gal gene expression. These data strongly

support the contention that the b-gal protein detected in In summary, adenoviral-microsphere complexes offer

Nahman et al: Microsphere-adenoviral complexes and transduction1510

Kronenberger J, Sedmak DD: Successful DNA transfer in cul-a simple and efficient approach to glomerular cell trans-tured human mesangial cells using replication deficient recombi-

duction. These studies demonstrate the utility of a micro- nant adenovirus. J Invest Med 46:204–209, 199813. Kim S, Glick A, Sporn M, Roberts A: Characterization of thesphere-based system for targeted gene delivery to the

promoter region of the human transforming growth factor B1 gene.glomerulus. Technical advances with biodegradable par-J Biol Chem 264:402–408, 1989

ticles and novel adenoviruses give this approach poten- 14. Sferra T, McNeely D, Johnson P: Gene transfer to the intestinaltract: A new approach using selective injection of the superiortial for eventual gene therapy of glomerular disease.mesenteric artery. Hum Gene Ther 8:681–687, 1997

15. Nahman NS Jr, Leonhart KL, Cosio FG, Hebert CL: Effects ofACKNOWLEDGMENTS high glucose on cellular proliferation and fibronectin production

by cultured human mesangial cells. Kidney Int 41:396–402, 1992This study was supported by National Institutes of Health grants16. Nahman NS Jr, Rothe K, Falkenhain M, Frazer KM, Dacio1R01 DK55115-01 (N.S.N.) and 1R21 DK57223-01 (N.S.N.), and grants

LE, Madia JD, Leonhart KL, Kronenberger JC, Stauch DA:from the American Diabetes Association (N.S.N.), the National Kid-Angiotensin II induction of fibronectin biosynthesis in culturedney Foundation of Ohio (N.S.N.), the American Heart Association,human mesangial cells: Association with CREB transcription fac-Ohio Valley Affiliate #CO-97-24-S (N.S.N.), The William H. Davistor activation. J Lab Clin Med 127:599–611, 1996Scholarship Fund of the Ohio State University (N.S.N.), the American

17. Chomczynski P, Sacchi N: Single-step method of RNA isolation byHeart Association, Ohio Valley Affiliate #Co-95-11-1 (BJH), the Ameri-acid guanidinium thiocyanate-phenol-chloroform extraction. Analcan Heart Association, Ohio Valley Affiliate Student Research Fellow-Biochem 163:156–159, 1987ship #9804026 (A.E.T.), and a Battelle Summer Research Internship

18. Nuovo GJ: Histologic distribution of hepatitis A, B, C, D, E, and(K.E.U.). We thank Mr. Arthur Weeks and Ms. Kerry Turpening forG with concomitant cytokine response in liver tissue. Diagn Moltheir assistance with the digital imaging used in this article.Pathol 7:267–275, 1998

19. Nuovo G: PCR In Situ Hybridization: Protocols and ApplicationsReprint requests to Dr. N. Stanley Nahman, Jr., Division of Nephrol-(3rd ed). New York, Lippincott-Raven Press, 1997ogy, N-210 Means Hall, 1654 Upham Drive, The Ohio State University,

20. Moullier P, Griedlander G, Calise D, Ronco P, PerricaudetColumbus, Ohio 43210, USA.M, Ferry N: Adenoviral-mediated gene transfer to renal tubularE-mail:nahman-1@medctr.osu.educells in vivo. Kidney Int 45:1220–1225, 1994

21. Zhu G, Nicolson A, Cowley B, Rosen S, Sukhatme V: In vivoadenovirus-mediated gene transfer into normal and cystic rat kid-REFERENCESneys. Gene Ther 3:298–304, 1996

1. Striker L, Doi T, Elliot S, Striker G: The contribution of glomer- 22. Zhu G, Nicolson A, Zheng X, Strom T, Sukhatme V: Adenovirus-ular mesangial cells to progressive glomerulosclerosis. Semin Neph- mediated B-galactosidase gene delivery to the liver leads to proteinrol 9:318–328, 1989 deposition in kidney glomeruli. Kidney Int 52:992–999, 1997

2. Border W, Okuda S, Nakamura T: Extracellular matrix and glo- 23. Leopold PL, Ferris B, Grinberg I, Worgall S, Hackett NR,merular disease. Semin Nephrol 9:307–317, 1989 Crystal RG: Fluorescent virions: Dynamic tracking of the pathway

3. Heikkila P, Parpale T, Lukkarinen O, Weber M, Tryggvason of adenoviral gene transfer vectors in living cells. Hum Gene TherK: Adenovirus-mediated gene transfer into kidney glomeruli using 9:367–378, 1998an ex vivo and in vivo kidney perfusion system: First steps towards 24. Smith A: Viral vectors in gene therapy. Annu Rev Microbiol 49:gene therapy of Alport syndrome. Gene Ther 3:21–27, 1996 807–838, 1995

4. Tomita N, Higaki H, Morishita R, Kato K, Mikami H, Kaneda 25. Parks RJ, Chen L, Anton M, Sankar U, Rudnicki MA, GrahamY, Ogihara T: Direct in vivo gene introduction into rat kidney. FL: A helper-dependent adenovirus vector system: Removal ofBiochem Biophys Res Commun 186:129–134, 1992 helper virus by Cre-mediated excision of the viral packaging signal.

5. Akagi Y, Isaka Y, Arai M, Kaneko T, Takenaka M, Moriyama Proc Natl Acad Sci USA 93:13565–13570, 1996T, Kaneda Y, Ando A, Orita Y, Kamada T, Ueda N, Imai E: 26. Clark K, Sferra T, Johnson P: Recombinant adeno-associatedInhibition of TGF-b1 expression by antisense oligonucleotides sup- viral vectors mediate long-term transgene expression in muscle.pressed extracellular matrix accumulation in experimental glomer- Hum Gene Ther 8:659–669, 1997ulonephritis. Kidney Int 50:148–155, 1996 27. Yang Y, Nunes F, Berencsi K, Furth E, Gonczol E, Wilson J:

6. Maeshima Y, Kashihara N, Yasuda T, Sugiyama H, Sekikawa Cellular immunity to viral antigens limits E1-deleted adenovirusesT, Okamoto K, Kanao K, Watanabe Y, Kanwar YS, Makino H: for gene therapy. Proc Natl Acad Sci USA 91:4407–4411, 1994Inhibition of mesangial cell proliferation by E2F decoy oligodeoxy- 28. Morsy MA, Gu M, Motzel S, Zhao J, Lin J, Su Q, Allen H,nucleotide in vitro and in vivo. J Clin Invest 101:2589–2597, 1998 Franlin L, Parks RJ, Graham FL, Kochanek S, Bett AJ, Caskey

7. Imai E, Isaka Y, Akagi Y, Kaneda Y: Gene transfer into the CT: An adenoviral vector deleted for all viral coding sequencesglomerulus by the hemagglutinating virus of Japan-liposome method. results in enhanced safety and extended expression of a leptinExp Nephrol 5:112–117, 1997 transgene. Proc Natl Acad Sci USA 95:7866–7871, 1998

8. Kitamura M, Taylor S, Unwin R, Burton S, Shimizu F, Fine L: 29. Willmott N, Cummings J, Stuart J, Florence A: Adriamycin-Gene transfer into the rat renal glomerulus via a mesangial cell vec- loaded albumin microspheres: In-vivo distribution and drug releasetor: Site-specific delivery, in situ amplification, and sustained expres- rate in the rat. Biopharm Frug Dispos 6:91–104, 1985sion of an exogenous gene in vivo. J Clin Invest 94:497–505, 1994 30. Willmott N, Chen Y, Goldberg J, Mcardle C, Florence A:

9. Kitamura M: Gene delivery into the glomerulus via mesangial cell Biodegradation rate of embolized protein microspheres in lung,vectors. Exp Nephrol 5:118–125, 1997 liver and kidney of rats. J Pharm Pharmacol 41:433–438, 1989

10. Tanner G: Effects of kidney tubule obstruction on glomerular 31. Schwendeman S, Costantino H, Gupta R, Langer R: Peptide,filtration in rats. Am J Physiol 6:F379–F385, 1979 protein and vaccine delivery from implantable polymeric systems:

11. Wilson J: Molecular medicine: Adenoviruses as gene-delivery ve- progress and challenges, in Controlled Drug Delivery: Challengeshicles. N Engl J Med 334:1185–1187, 1996 and Strategies, edited by Park K, Washington D.C., American

12. Nahman NS Jr, Clark K, Sferra TJ, Urban KE, Troike AE, Chemical Society, 1997, pp 229–263