Em

JNIa

b

c

d

e

f

g

a

ARRA

KCECMMM

I

r(e

cAutchHa

D

a

1d

International Journal of Medical Microbiology 300 (2010) 449–456

Contents lists available at ScienceDirect

International Journal of Medical Microbiology

journa l homepage: www.e lsev ier .de / i jmm

nterobacterial tumor colonization in mice depends on bacterial metabolism andacrophages but is independent of chemotaxis and motility

ochen Stritzkera,b,c, Stephanie Weibelb,c, Carolin Seubertb,c, Andreas Götzb, Achim Treschd,ico van Rooijene, Tobias A. Oelschlaegerc, Philip J. Hill f,

vaylo Gentscheva,b,c, Aladar A. Szalaya,b,c,g,∗

Genelux Corporation, 3030 Bunker Hill Street, San Diego, CA 92109-5754, USADepartment of Microbiology and Institute of Biochemistry, Biocenter, University Würzburg, Würzburg, GermanyInstitute for Molecular Infection Biology, University Würzburg, Würzburg, GermanyGene Center, Department of Chemistry and Biochemistry, Ludwig-Maximilians-University, Munich, GermanyDepartment of Cell Biology and Immunology, Vrije Universiteit of Amsterdam, Amsterdam, The NetherlandsUniversity of Nottingham, School of Biosciences, Sutton Bonington Campus, UKDepartment of Radiation Oncology, Moores Cancer Center, University of California, San Diego, CA 92093-0843, USA

r t i c l e i n f o

rticle history:eceived 2 October 2009eceived in revised form 14 April 2010ccepted 14 April 2010

a b s t r a c t

Despite promising results and increasing attention in bacterial cancer therapy, surprisingly little is knownabout initial tumor colonization and the interaction between bacteria and surrounding tumor tissue. Here,we analyzed the role of chemotaxis, motility, and metabolism both in Escherichia coli and Salmonellaenterica serovar Typhimurium strains upon intravenous injection into tumor-bearing mice. In contrastto previous models, we found that chemotaxis and motility do not play a significant role in tumor col-

eywords:ancernterobacteriahemotaxisacrophageetabolism

onization and bacterial distribution within the tumor. Rather, the whole colonization and intratumoralmigration process seems to be a passive mechanism that is influenced by the reticuloendothelial sys-tem of the host, by the tumor microenvironment and by the bacterial metabolism. These conclusionswere supported by experimental data demonstrating that disruption of the basic branch of the aromaticamino acid biosynthetic pathway and depletion of macrophages, in contrast to flagellar mutations, led

bacte

echanism of action to significant changes in

ntroduction

In the past several years, cancer therapy with live bacteria hasegained attention and significant advances have been achievedLoeffler et al., 2007; Mengesha et al., 2007; Royo et al., 2007; Weit al., 2007; Zhang et al., 2007).

Initial studies using intentional bacterial injection into humanancer patients had already been carried out in 1866 by W. Busch.t the end of the same century, F. Fehleisen and, later on, W. Coleysed live Streptococcus pyogenes and Serratia marcescens as tumor-herapy agents (Kienle and Kiene, 2003). W. Coley, who is now

redited as the founder of cancer immunotherapy, then changedis therapeutic approach to cell extracts of the same bacteria.owever, due to the success of radio- and chemotherapies, andbetter understanding of the principles behind these therapies,

∗ Corresponding author at: Genelux Corporation, 3030 Bunker Hill Street, Saniego, CA 92109-5754, USA. Tel.: +1 909 307 9300; fax: +1 909 307 2251.

E-mail addresses: aaszalay@genelux.com,ladar.szalay@virchow.uni-wuerzburg.de (A.A. Szalay).

438-4221/$ – see front matter © 2010 Elsevier GmbH. All rights reserved.oi:10.1016/j.ijmm.2010.02.004

rial accumulation in tumors of live mice.© 2010 Elsevier GmbH. All rights reserved.

the medical community has preferred them to bacterium-mediatedtherapies. In the middle of the 20th century, scientists again becameinterested in using bacteria, especially Clostridium spp., as ther-apeutic agents and, meanwhile, several studies were performedusing genetically engineered bacterial species, including virulence-attenuated mutants of pathogenic species that could be injectedinto mouse tumor models without causing disease, and/or express-ing newly added heterologous therapeutic genes (for review seeWei et al., 2007). The use of Salmonella spp. has been particularlysuccessful in many murine tumor models (Pawelek et al., 2003) andthe VNP20009 strain has even been tested in clinical trials in humancancer patients (Toso et al., 2002).

Despite the regained attention in bacterial cancer therapy,surprisingly little is known about initial tumor colonization andthe interaction between bacteria of the Enterobacteriaceae fam-ily and surrounding tumor tissue. Here, we tried to shed light on

these events by analyzing the role of chemotaxis, motility, andmetabolism, both in E. coli and S. typhimurium strains, as well asthe effects of systemic macrophage depletion. In contrast to pre-vious models, our data described in this manuscript indicate thatactive chemotaxis and motility did not play a significant role in

450 J. Stritzker et al. / International Journal of Medical Microbiology 300 (2010) 449–456

Table 1Bacterial strains used in this study.

Strain Relevant characteristics Reference or source

E. coli Nissle 1917 Probiotic E. coli Altenhoefer et al. (2004)E. coli Nissle 1917 �fliA fliA::cat Schlee et al. (2007)E. coli Nissle 1917 �fliC fliC::cat Schlee et al. (2007)E. coli Nissle 1917 �flgE flgE::cat Schlee et al. (2007)E. coli Nissle 1917 �fliC pDB2 fliC::cat, fliC expressed in trans from pDB2 Schlee et al. (2007)E. coli Nissle 1917 �flgE pDB3 flgE::cat, flgE expressed in trans from pDB3 Schlee et al. (2007)E. coli Nissle 1917 �aroA (pENTR221-GFP) aroA::cat This studyE. coli Nissle 1917 �aroA pENTR221-aroA aroA::cat, aroA expressed in trans from pENTR221-aroA This studyE. coli 4608-58 Enteroinvasive E. coli Hale et al. (1983)E. coli 4608-58 �aroA (pENTR221-GFP) aroA deletion This studyE. coli 4608-58 �aroA (pENTR221-aroA) aroA deletion, aroA expressed in trans from pENTR221-aroA This study

TyphSL134L1344L1344

toow

M

B

i02iAoob

nod

(ib

a

um(CTCGGCcpialDTf

S. typhimurium SL1344 Virulent S. enterica serovarS. typhimurium M913 fliGHI::Tn10, derived fromS. typhimurium M935 cheY::Tn10, derived from SS. typhimurium SL7207 aroA::Tn10, derived from S

umor colonization or intratumoral migration, whereas disruptionf the aromatic amino acid biosynthetic pathway and depletionf macrophages significantly altered the bacterial accumulationithin tumors of mice.

aterials and methods

acteria

All bacterial strains used in this study are listed in Table 1.For injection of tumor-bearing animals, bacteria were grown

n LB broth at 37 ◦C and 190 rpm until reaching an OD600 nm of.4, which corresponds to about 2 × 108 CFU/ml (Stritzker et al.,007). Bacteria were harvested by centrifugation and washed twice

n endotoxin-free phosphate buffered saline (PBS) (PAA, Pasching,ustria). The bacteria were then diluted accordingly and 100 �lf the suspension (containing approximately 5 × 106 CFU if nottherwise indicated) was injected into the lateral tail vein of tumor-earing mice.

The number of bacteria as CFU/g was obtained after homoge-ization of respective tissue in 1.0 ml ice-cold PBS and plating 0.1 mlf serial dilutions of the suspension on LB agar plates. The limit ofetection was about 100 CFU/g tissue analyzed.

Determination of bacterial load in tumors was done at 6 hmacrophage depletion experiments, confocal microscopy stud-es) or 48 h (chemotaxis and motility studies, aromatic amino acidiosynthesis mutants) post-injection.

roA deletion and complementation

Strains E. coli 4608-58 (EIEC) and E. coli Nissle 1917 (EcN) weresed for construction of the respective aroA mutants using theethod developed by Datsenko and Wanner (2000). Primer pairs

5′-TTT TAT TTC TGT TGT AGA GAG TTG AGT TCA TGG AAT CCC TGAGG TGT AGG CTG GAG CTG CTT C-3′)/(5′-AGA TTT GGC TAT TTATG CCC GTT GTT CAT TCA GGC TGC CTG GCT CAT ATG AAT ATCTC CTT A-3′) for EIEC and (5′-TGG GGT TTT TAT TTC TGT TGT AGAAG TTG AGT TCG TGT AGG CTG GAG CTG CTT-3′)/(5′-AGA AAGAT TGT CTA TGT TAT CGC CCG TTA TTC ACA TAT GAA TAT CCTCT TAG TTC CTA-3′) for EcN, respectively, were used to amplify thehloramphenicol resistance cassette (cat) from pKD3. After electro-oration of the PCR product, transformants were cultured for 1 day

n LB supplemented with 10 �g/ml vitamin K2 and 0.2% l-arabinose

nd then plated on LB agar plates supplemented with 0.2% (w/v)-arabinose, 30 �g/ml chloramphenicol and 10 �g/ml vitamin K2.eletion of aroA was confirmed by PCR using the primer pairs (5′-TA TAC GCA AGG CGA CAA GG-3′/5′-CAG TTG GCG GAC AGT G-3′)or EIEC 4608-58 or (5′-CAG CAT AAT CCC CAC AGC CA-3′) (5′-CAC

imurium ‘wild-type’, histidine auxotroph B.A.D. Stocker4 Stecher et al. (2004)

Stecher et al. (2004)Hoiseth and Stocker (1981)

AAG GTC CGA AAA AAA ACG C-3′) for EcN. In EIEC, the resistancecassette was additionally removed by transforming the aroA::catmutant with pCP20 (Cherepanov and Wackernagel, 1995). Deletionof the chloramphenicol resistance gene was confirmed by PCR.

For in trans complementation of aroA mutants, the aroA geneincluding its promoter region was PCR amplified from wild-typeEcN using the primer pair 5′-GGG GAC AAG TTT GTA CAA AAAAGC AGG CTA AGG AGG AAT AAA AAG CCA TGC CGC TGG AAGGTG T-3′/5′-GGG GAC CAC TTT GTA CAA GAA AGC TGG GTC CGGCAA TGT GCC GAC GTC TT-3′ which also contained attB1 and attB2sites. The resulting PCR fragment was cloned into pDONRTM221(Invitrogen, Carlsbad, CA, USA) using BP-clonase according to themanufacturer’s instructions. The resulting pENTR221-aroA wasthen electroporated into the aroA mutants of EcN and EIEC 4608-58.Controls were transformed with pENTR221-GFP.

Soft agar motility assay

Testing of bacterial motility was performed in a soft agar motil-ity assay. Bacteria were stabbed into semisolid LB agar plates with0.25% agar and incubated at 37 ◦C for 6 (S. typhimurium) or 8(EcN) hours. The diameter of the bacteria-containing turbid regionaround the inoculation site was determined using an electroniccaliper and 1 mm was subtracted (diameter of the tip used to stabthe bacteria into the agar). Data were given as % motility of eachwild-type bacterial strain (note that S. typhimurium has about 3-foldhigher motility than EcN in the soft agar motility assay).

Histology and fluorescence microscopy

Histology and fluorescence microscopy were performed asdescribed previously (Stritzker et al., 2007; Weibel et al., 2008).Briefly, snap-frozen tumors were fixed in 4% formaldehyde, and100-�m vibratome tissue sections were permeabilized in PBS con-taining 0.3% Triton X-100. Sections were then incubated withFITC-labeled phalloidin or rat anti-mouse CD68 antibodies (Serotec,Düsseldorf, Germany), Hoechst 33342 (optional) and biotinylatedpolyclonal antibodies against E. coli or Salmonella spp. (ViroStat,Portland, USA) for 12–15 h, followed by washing and incubationwith Cy3-conjugated streptavidin (Sigma, Taufkirchen, Germany)and, when rat anti-mouse CD68 primary antibodies were used, Cy3-conjugated donkey anti-rat IgG antibodies. After several rinses inPBS, tissue sections were incubated in PBS containing 60% (v/v)

glycerol and then mounted in PBS containing 80% (v/v) glycerol.

Confocal microscopy was performed using a Leica TCS SP2 AOBSequipped with an argon, helium–neon and UV laser. Digital imageswere processed with Photoshop 5.0 (Adobe Systems, MountainView, USA) and merged to yield pseudo-colored images.

J. Stritzker et al. / International Journal of Medical Microbiology 300 (2010) 449–456 451

Fig. 1. Tumor colonization of wild-type E. coli Nissle 1917 and motility mutants. (A) 4T1 tumor-bearing BALB/c mice were intravenously injected with 1 × 106 CFU of E.coli Nissle 1917 wild-type (wt), flagella mutants (�fliA, �fliC, �flgE), or in trans complemented mutants (�fliC pDB2, �flgE pDB3). The motility of the different strains wast dardt rmine( issle 1c

P

mabibfFI

C

c

ascPta

pC‘

M

swGafGcibs

ested in a soft agar motility assay (grey bars showing the average values plus stanrans complemented mutants). Furthermore, the number of CFU/g tumor was deteB) Sections (100 �m) of formaldehyde-fixed tumors of mice injected with E. coli Noli Nissle 1917 (red). Scale bars indicate 5.0 mm.

eritoneal lavage and flow cytometry

For peritoneal lavages, the peritoneal cavities of sacrificedice were washed with 3 ml PBS containing 3 mM ethylendi-

minetetraacetate (Ajuebor et al., 1998). After washing with Hanks’alanced salt solution containing 1% fetal calf serum, cells were

ncubated with PE-labeled anti-CD45 (Beckton Dickinson, Heidel-erg, Germany) and FITC-labeled anti-F4/80 (Beckton Dickinson)or 30 min at 4 ◦C, and after washing analyzed on a FACSCalibur.or data interpretation, FlowJo software (Version 7.2.5, Tree Starnc., Ashland, USA) was used.

ell culture and animal models

4T1 cells were cultured in RPMI supplemented with 10% fetalalf serum at 37 ◦C in a 5% CO2 containing atmosphere.

BALB/c mice were obtained from Harlan (Indianapolis, USA,nd Borchen, Germany). Five- to six-week-old female mice wereubcutaneously injected with 3.3 × 104 murine 4T1 mammary can-er cells (ATCC: CRL-2539) resuspended in 100 �l endotoxin-freeBS. Intravenous injection of bacteria was performed 14 days afterumor cell implantation. At this time the tumors had a diameter ofbout 1 cm.

All animal experiments were carried out in accordance withrotocols approved by the Institutional Animal Care and Useommittee (IACUC) of Explora Biolabs (San Diego, USA) or the

Regierung von Unterfranken’ (Würzburg, Germany).

acrophage depletion

For macrophage depletion, a suspension of clodronate lipo-omes (CLIP) containing about 6 mg clodronate per 1 ml suspensionas used. Clodronate was a gift of Roche Diagnostics (RochembH, Mannheim, Germany) and it was encapsulated in liposomess described earlier (Van Rooijen and Sanders, 1994). Reagentsor preparation of liposomes were: phosphatidylcholine (Lipoid

mBH, Ludwigshafen, Germany) and cholesterol (Sigma Chemi-als Co., Germany). Depletion of macrophages was achieved byntraperitoneal (i.p.) injection of 0.2 ml CLIP 2 days and 4 h beforeacterial injection. Control mice were injected with 0.2 ml lipo-omes containing PBS.

deviation; n ≥ 10; *p < 0.001 compared to wt, and **p < 0.001 compared to wt or ind 2 days post-injection (black bars, average values plus standard deviation; n = 4).917 wt or E. coli Nissle �fliA, �fliC, and �flgE were stained for actin (green), and E.

Statistical analysis

Data were analyzed using Wilcoxon rank sum test. A p value ofless than 0.05 was considered statistically significant.

Results

Tumor colonization by motility mutants of E. coli Nissle 1917

We recently showed that E. coli Nissle 1917 (EcN) is able to colo-nize, replicate and survive for at least 25 days within 4T1 syngeneicbreast tumors in immunocompetent BALB/c mice with tumor toorgan ratios of >106:1 (Stritzker et al., 2007). Here, we set out toinvestigate the steps that are responsible for successful tumor colo-nization. A recent publication showed that chemotaxis and motilityare crucial for efficient colonization of Salmonella typhimurium ina tumor cylindroid model in vitro (Kasinskas and Forbes, 2007).Here, we initiated experiments to verify these results in vivo intumor-bearing animals.

Therefore, we analyzed wild-type (wt) and motility mutantstrains of EcN (Schlee et al., 2007). While wt EcN was able to spreadinto soft agar plates, �fliA (deficiency for the sigma factor respon-sible for flagellum gene synthesis), �fliC (deficient in the flagellinfilament protein), and �flgE (absence of essential flagella hook pro-tein) mutants were not able to actively move into the surroundingsoft agar (Fig. 1A, grey bars). The motility deficiency in the �fliCand �flgE mutants could be overcome when flgE and fliC wereexpressed in trans. Plasmid-based expression of flgE (from pDB3)and fliC (from pDB2) led to 88.6% and 83.7% motility, respectively,when compared to wt EcN.

Surprisingly, when wt and motility-deficient EcN bacteria wereintravenously (i.v.) injected into 4T1 tumor-bearing BALB/c mice,no difference in terms of tumor colonization ability was observed(Fig. 1A, black bars). The deficiency in motility, therefore, did nothave any effect on the initial colonization or the replication of intra-tumoral EcN.

We then tested whether motility-deficient mutants of EcN at

least showed different distribution patterns compared to wt EcN.However, as shown in Fig. 1B, the patterns obtained from wt andmotility mutant colonization experiments were not significantlydifferent from each other. All strains tested seemed to form patchesin the necrotic tumor center and accumulated next to the previ-

452 J. Stritzker et al. / International Journal of Medical Microbiology 300 (2010) 449–456

Fig. 2. Tumor colonization of wild-type Salmonella enterica serovar Typhimurium, chemotaxis- and motility mutants. (A) The motility of different S. typhimurium strains wast FU/g tp 01 comfi uriums

otat

Te

pEtdab(pcs

mo

ftrno

ested in a soft agar motility assay (left panel; n = 6). Furthermore, the number of Canel, n = 4). Data represent average plus standard deviation with * indicating p < 0.0xed tumors of mice colonized with wild-type S. typhimurium SL1344 (wt), S. typhimtained for actin (green), and Salmonella spp. (red). Scale bars indicate 5.0 mm.

usly defined border region between the necrotic and live tumorissue. This border region was previously characterized by elevatedctin concentrations and the accumulation of inflammatory infil-rates (Weibel et al., 2008; Westphal et al., 2008).

umor colonization by chemotaxis- and motility mutants of S.nterica serovar Typhimurium

To investigate whether chemotaxis and motility are also dis-ensable for tumor colonization of S. typhimurium, and not only. coli, in live animals, S. typhimurium wt (SL1344) was comparedo chemotaxis- (cheY−; M935 – unable to detect chemical gra-ients but still encoding for all genes responsible for motility)nd motility- (flgGHI−; M913 – able to detect chemical gradientsut unable to produce functional flagella) deficient mutant strainsStecher et al., 2004). In the soft agar motility assay, neither mutantroved to be significantly motile (Fig. 2A, left panel), although theheY-deficient strain seemed to be able to at least penetrate theurrounding soft agar.

As observed for EcN, neither the use of the chemotaxis- nor theotility-deficient mutant resulted in significant changes in terms

f tumor or organ colonization (Fig. 2A, right panel).Furthermore, histological analysis did not reveal significant dif-

erences among wild-type, chemotaxis-, or motility mutants of S.yphimurium (Fig. 2B). All three strains colonized both the necroticegions and accumulated at the border region between live andecrotic tumor tissue. The results are very similar to those alreadybtained for the motility-deficient mutants of EcN, although the

umor, liver and spleen was determined 2 days post-injection of 1 × 106 CFU (rightpared to wild-type S. typhimurium SL1344. (B) Sections (100 �m) of formaldehyde-M935 cheY mutant (cheY−), or S. typhimurium M913 fliGHI mutant (fliGHI−) were

border region in tumors colonized with the S. typhimurium strainsseemed to be closer to the tumor edge in comparison to thosecolonized with E. coli.

Depletion of macrophages resulted in enhanced tumorcolonization by E. coli Nissle 1917 in contrast to S. entericaserovar Typhimurium

Apart from bacterial factors needed for efficient and specifictumor colonization, we also investigated the role of host mono-cytes and macrophages in this process. One may expect thatmacrophages and monocytes take up bacteria from the blood-stream and these cells can subsequently be recruited to the tumor.Once inside the tumor tissue, the bacteria might be able to escapefrom the cells. Since Salmonella infections are mediated in a simi-lar way by macrophages, this process is a feasible mechanism fortumor colonization (Sansonetti and Phalipon, 1999). For EcN as non-pathogenic bacteria, this route is less likely, but it is conceivablethat EcN could use this route since the tumor environment causesmacrophages to change their properties from M1- to M2-typetumor-associated macrophages, which are less efficient effectorcells (Mantovani et al., 2004) and might allow bacterial escape.On the other hand, depletion of phagocytic cells might prolong

circulation time and survival within injected animals.

To investigate the role of macrophages and monocytes in ini-tial tumor colonization, clodronate containing liposomes (CLIP)were i.p. injected, which leads to depletion of phagocytic cells(Van Rooijen and Sanders, 1994). PBS-containing liposomes were

J. Stritzker et al. / International Journal of Med

Table 2Percentage of F4/80-positive macrophages and monocytes in the CD45+ lymphocytepopulation.

Monocytes(F4/80low)

Macrophages(F4/80high)

Total (F4/80pos.)

Control mice 41.8 ± 9.9% 6.4 ± 3.2% 48.2 ± 10.4%

D

iiEa

dlWt

emsnmcdbsd

FnshflPrdwsa

CLIP-treated mice 5.2 ± 1.7% 0.2 ± 0.1% 5.3 ± 1.7%

ata indicate average ± standard deviation (n = 8).

njected as a control. As initial events after tumor colonization werenvestigated, mice were sacrificed 6 h after injection of SL1344 orcN. Then a peritoneal lavage was performed, and tumors, spleennd liver were isolated.

The peritoneal lavages were used to confirm macrophageepletion by analyzing the amount of F4/80+ cells in the CD45+

ymphocyte population using flow cytometry (Table 2 and Fig. 3A).hile PBS-liposome-injected mice contained 48.2 ± 10.4%, CLIP-

reated mice only had 5.3 ± 1.7% F4/80+ lymphocytes.Analysis of bacterial load in liver, spleen and tumors showed

levated numbers of EcN bacteria in tumors of CLIP-treatedice, compared to PBS-liposome-treated mice (Fig. 3B and

upplementary Fig. S1). While macrophage depletion did not sig-ificantly affect the colonization of tumors in SL1344-injectedice (p > 0.28) 6 h after injection, the number of intratumoral EcN

hanged profoundly (p < 0.002). In livers, macrophage depletion

id not result in significant changes of bacterial load (p > 0.5 foroth SL1344 and EcN), whereas the number of bacteria found inpleens increased (p < 0.02 for SL1344 and p < 0.001 for EcN). Theseata indicate that monocytes and macrophages did affect success-

ig. 3. Depletion of macrophages affects tissue distribution and tumor colo-ization. (A) Peritoneal lavages of PBS-liposomes- and CLIP-treated mice weretained for CD45 and F4/80. The percentage of non-, low- (monocytes), andigh- (macrophages) F4/80-expressing lymphocytes (CD45+) was determined byow cytometry. One representative example of each group is depicted. (B)BS-liposomes- and CLIP-treated mice were i.v. injected with EcN and SL1344,espectively. Presence of bacteria was analyzed 6 h post-injection by plating serialilutions of tumor, liver and spleen lysates on LB agar plates. The number of CFU/gas calculated and average + standard deviation from at least 6 mice per group is

hown. Asterisk indicates statistical significance (p < 0.02) when comparing CLIP-nd PBS-liposomes-treated mice injected with the same bacterial strain.

ical Microbiology 300 (2010) 449–456 453

ful tumor colonization, in case of EcN, but were not involved whenSL1344 strain was used to colonize tumor xenografts in mice.

Tumor microenvironment during S. enterica serovarTyphimurium colonization

Although macrophage depletion did not abrogate the accu-mulation of SL1344 in tumors, we investigated whether tumor-associated macrophages (not efficiently depleted by CLIP, asobserved in CD68 stained tumor sections – not shown) or othercells were invaded by Salmonella. Therefore, confocal microscopyof phalloidin- or anti-CD68-stained tumor sections was performedat 6 and 48 h post-injection (p.i.) of SL1344.

Although, due to the low number of bacteria in tumors at 6 hp.i., it was difficult (and impossible for EcN) to confidently ana-lyze the distribution of bacteria in tumor sections, we were ableto examine a number of SL1344 microcolonies within these sec-tions. While at 6 h p.i. we could detect S. typhimurium both intra-and extracellularly (Fig. 4A), most of the bacteria were outsidecells at 48 h post-injection (data not shown). We were not able toshow co-localization of anti-CD68 stained cells with SL1344 at 6 hp.i., although we could detect bacteria in close proximity to CD68-positive cells and co-localization with cell fragments, respectively(Fig. 4B, and supplementary Fig. S2).

Tumor-selective colonization of aromatic amino acidbiosynthesis-deficient bacteria

After excluding a significant role of chemotaxis and motilityduring tumor colonization in live animals, the role of bacterialmetabolism was investigated. In this study, we focused on thebasic branch of the aromatic amino acid biosynthesis pathway.Disruption of this pathway is frequently used to generate virulence-attenuated strains of pathogenic bacteria (e.g. Bacillus anthracis(Ivins et al., 1990), Bordetella pertussis (Roberts et al., 1990), Lis-teria monocytogenes (Stritzker et al., 2004), Neisseria gonorrhoeae(Chamberlain et al., 1993), Pseudomonas aeruginosa (Priebe et al.,2002), Shigella flexneri (Kotloff et al., 1996), Staphylococcus aureus(Buzzola et al., 2006), Salmonella typhi (Dougan et al., 1987), andSalmonella typhimurium (Hoiseth and Stocker, 1981)), which canthen be used as vaccine carrier strains against autologous, as wellas heterologous antigens. In terms of cancer therapy and/or can-cer diagnosis, the use of non-virulent and/or virulence-attenuatedstrains is highly desirable to prevent problems caused by infectionsof secondary sites other than the tumor.

When 4T1 tumor-bearing mice were injected with 5 × 106 CFUof either wild-type or aroA-deficient S. typhimurium SL7207, EIECor EcN, the resulting organ colonization pattern 2 days post-injection changed dramatically (Fig. 5A). While the aroA mutantstrain SL7207 of S. typhimurium reached about 10% of the wild-type level in tumors, the corresponding mutation in EIEC and EcNstrains led to about 2500-fold and 200-fold reductions in tumorcolonization, respectively. In comparison, bacterial load in liverswas drastically reduced for both aroA-deficient pathogen-derivedstrains (about 4.5 × 103-fold for S. typhimurium and 2.0 × 104-foldfor EIEC), when compared to their corresponding wild-type strains.Moreover, SL7207 colonization of the spleen was found to givealmost a 3.0 × 104-fold reduction, in comparison to the SL1344Salmonella wild-type strain, while none of the E. coli strains weredetected in spleens of 4T1 tumor-bearing mice.

To exclude the possibility that the effects were caused by

secondary mutations, the aroA mutants of the E. coli strainswere complemented in trans with the aroA-encoding plasmidpENTR221-aroA or pENTR221-GFP (as negative control). The result-ing strains and wild-type EcN and EIEC 4608-58 were i.v. injectedinto 4T1 tumor-bearing BALB/c mice (5 × 105 CFU) and tumor colo-

454 J. Stritzker et al. / International Journal of Medical Microbiology 300 (2010) 449–456

F croscoI are sh

nirsa

D

ie(2(obtC(at2puocge

c

ig. 4. Sections of tumors at 6 h post-injection of SL1344 analyzed by confocal mintracellular bacteria in (A) appear yellow and are indicated by arrows. Two panels

ization was analyzed 2 days post-injection (Fig. 5B). The reductionn colonization caused by the genomic aroA deletion could be fullyestored by the presence of aroA-encoding plasmid, proving thatecondary mutations played no role in tumor colonization by theroA mutants.

iscussion

Enterobacterium-mediated tumor-therapy strategies – includ-ng the use of unmodified bacteria alone (Low et al., 2004; Zhaot al., 2007), modified bacteria with prodrug-converting enzymesPawelek et al., 1997; Royo et al., 2007), siRNAs (Zhang et al.,007), cytokines (Loeffler et al., 2007), anti-angiogenic proteinsLee et al., 2004, 2005a), and bacteria in combination with radio-r chemotherapy (Platt et al., 2000; Lee et al., 2005b) – recentlyecame available. However, requirements for successful, specificumor colonization still remain unresolved and little understood.urrently, the uneven distribution of the enterobacteria around thebacteria-induced) necrotic center of tumors 2 days post-injection,ppears to be the result of the innate immune system leadingo ‘encapsulation’ and containment of the bacteria (Weibel et al.,008; Westphal et al., 2008). At this time (2 days p.i.), bacterialromoters activated inside tumors were recently characterized bysing a random library of S. typhimurium genomic DNA upstreamf a promoterless GFP, in combination with fluorescence-activated

ell sorting and oligonucleotide tiling arrays (Arrach et al., 2008)enerally showing the upregulation of hypoxia-inducible promot-rs.

In contrast to the above experiments, early events during tumorolonization are not well studied, apart from a recent investigation

py. Actin (A) and CD68 (B) are shown in green, nuclei in blue and SL1344 in red.own for each staining.

on Salmonella choleraesuis showing decreased tumor colonizationability upon presence of bacteria-specific antibodies (Lee et al.,2009). The role of specific bacterial and tumor cell genes are alsounknown during these events. For strict anaerobes like clostridia orbifidobacteria, anoxic necrotic centers within the tumor are essen-tial for successful tumor colonization. The requirements and needsfor facultative aerobic enterobacteria for tumor colonization, how-ever, are less clear. Recently, Kasinskas and Forbes (2007) reportedthe necessity of chemotaxis for successful bacterial initiation, pen-etration and colonization of a tumor cylindroid model in vitro. Inparticular, these authors also suggested that chemotaxis is essentialfor tumor colonization in vivo. Further, they claimed that the dele-tion of the ribose/galactose transporter allowed the developmentof improved targeting of bacterial therapies to otherwise inacces-sible regions of tumors. However, here we clearly demonstratedthat bacterial tumor colonization and intratumoral distribution areindependent of motility in live tumor-bearing mice, both for intra-venously injected EcN and S. typhimurium. As chemotaxis is motilitydependent we conclude that the tumor colonization process is alsoindependent of chemotaxis, which was also proven experimentallyusing a cheY mutant of S. typhimurium. In these experiments, theamount of bacteria within the tumor and their localization was ana-lyzed 48 h after injection as the bacterial distribution pattern at thistime is very consistent and it would have been easy to detect vari-ance from this pattern. During earlier stages of tumor colonization

the bacteria are more “randomly” distributed and so definitive dataare very difficult to collect.

We propose, therefore, that although the in vitro tumor cylin-droid model may be a valuable tool for testing of hypotheses andmaking predictions about the therapeutic efficacy of drugs and/or

J. Stritzker et al. / International Journal of Med

Fig. 5. Tumor colonization and tissue distribution of deletion mutants in the basicbranch of amino acid biosynthesis. (A) Deletion of aroA (striped bars) led to reducedcolonization of tumors but to higher tumor to organ ratios in pathogenic Salmonellatyphimurium and E. coli 4608-58. (B) The wild-type strains were compared to thearoA deletion mutants of EcN and E. coli 4608-58 in terms of tumor colonization. Themutant strains were either transformed with pENTR221-GFP (negative control) orpENTR221-aroA (in trans complementation). The number of CFU/g was calculatedam

mtettdpor

dmetrdpilrb

nd average + standard deviation is shown; * indicates p < 0.01 when comparing aroAutant to wild-type bacteria in the same organ.

icroorganisms, one has to be very careful when extrapolatinghese predictions for tumors in vivo. In live animals, the bacterianter the tumor via the vasculature and colonization occurs fromhe inside rather than from the outside, as is the case in the artificialumor cylindroid model. We could confirm that colonization andistribution are processes independent of motility as already pro-osed by R.K. Jain and colleagues (Forbes et al., 2003) and that thebserved distribution pattern results from a largely passive process,eflecting the tumor microenvironment.

These findings are also supported by macrophage depletionata. Indirect transport of the bacteria within monocytes oracrophages in the bloodstream to the tumors can probably be

xcluded since depletion experiments did not result in abroga-ion of tumor colonization. In contrast, depletion of macrophagesesulted in elevated numbers of EcN within tumor tissues earlyuring the tumor colonization process. This may occur due to

rolonged survival of bacteria in the bloodstream that are not elim-

nated by the reticuloendothelial system. Therefore, the chance ofarger numbers of bacteria passively entering the tumor microenvi-onment where they can start replication (e.g. in leaky or dead-endlood vessels) is enhanced. For EcN the number of injected bac-

ical Microbiology 300 (2010) 449–456 455

teria necessary to successfully colonize the tumor could thereforebe lower in the absence of macrophages. In the long run, however,the presence or absence of macrophages will not make a differ-ence since the number of bacteria will reach a maximum of about108 to 109 cfu/g regardless of the initial number of bacteria withinthe tumor. Using Salmonella, the tumor colonization is more or lessunaffected by macrophage depletion since these bacteria are fac-ultative intracellular and are able to survive and replicate insidethese cells.

On the other hand, S. typhimurium might use cell invasion asan entry port into tumors. At this time, we can neither excludenor support this assumption since we found both intracellular aswell as extracellular bacteria in tumors 6 h after intravenous injec-tion. The higher numbers of S. typhimurium in the tumor at thistime, compared to EcN, might be attributed to protection in invadedcells or to the higher serum sensitivity of the semirough EcN strain(Grozdanov et al., 2002).

At 48 h post-injection most of the bacteria were extracellular(data not shown), which was in line with previous publications andthe fact that enterobacterial strains can colonize tumors regardlessof their ability to replicate intracellularly (Stritzker et al., 2007).

Taken together, we suggest that, after the initial colonization,bacterial products (such as endotoxin, flagella, bacterial genomicDNA etc.) activate Toll-like receptors followed by host responsesthat lead to tumor tissue destruction, remodeling and infiltration ofmacrophages and/or neutrophils that enclose the bacteria (Weibelet al., 2008; Westphal et al., 2008; Leschner et al., 2009). The bac-teria replicate inside the necrotic tumor tissue and by internalinterstitial pressure (present in both human and experimental solidtumors (Jain, 1994, 1998; Heldin et al., 2004)), are passively trans-ported to the tumor rim where further spreading is restricted byphagocytic effector cells resulting in the observed ring structure.

Replication of the bacteria within the tumor (and also in othertissues) is, of course, dependent on intact bacterial metabolism. Bycomparing wild-type and aroA mutants of S. typhimurium, EIEC, andEcN we were able to show that the basic branch of aromatic aminoacid biosynthesis plays a crucial role not only in spleen and livercolonization (which is drastically reduced in each strain), but alsofor replication in tumors. Surprisingly, the aroA deletion has a muchstronger effect on the E. coli strains than on S. typhimurium, althoughboth S. typhimurium and EIEC are facultative intracellular bacterialstrains and EcN is an extracellular strain. In contrast to the E. colistrains, S. typhimurium is only marginally affected (6-fold reduc-tion in tumors) and the tumor to organ ratios improve from about40:1 and 180:1 for liver and spleen, respectively, to about 6500:1and 1750:1 in the same organs. As exemplified for the aroA geneof the basic branch of the aromatic biosynthesis, it would be inter-esting to further characterize the role of bacterial metabolism fortumor-specific replication and investigate the effects in strains withdeletions on other key components in basic metabolic pathwaysand directly compare them to wild-type strains. Such comparisonsof different metabolic pathways will help to explain the differ-ences between E. coli and S. typhimurium, and give reasons for whySalmonella, in contrast to E. coli, can replicate intracellularly andinduce therapeutic effects.

Taken together, the colonization of tumors by bacteria of theEnterobacteriaceae family seems to be a largely passive processthat is dependent on the metabolic properties of the bacteriabut independent of their active motility towards nutrient gradi-ents. Host factors do not seem to play a role after the bacteriareach the immunoprivileged ‘tumor sanctuary’ where the bacte-

ria seem to be protected from the immune system. On the otherhand, the immune system is activated during enhanced intratu-moral bacterial replication, and recruited phagocytic cells inhibitthe bacteria from spreading further into the remaining tumortissue.

4 of Med

A

neaAr

A

t

R

A

A

A

B

C

C

D

D

F

G

H

H

H

I

JJ

K

K

K

L

L

Typhimurium carrying plasmid-based small interfering RNAs. Cancer Res. 67,5859–5864.

56 J. Stritzker et al. / International Journal

cknowledgements

The authors thank J. Langbein and L. Kirscher for excellent tech-ical assistance, M. Adelfinger for animal care, A. Feathers forditorial help, and W.D. Hardt for providing S. typhimurium M913nd S. typhimurium M935. J. Stritzker, P.J. Hill, I. Gentschev, and.A. Szalay have financial interest in Genelux Corp., San Diego. Theesearch was supported by Genelux Corp., San Diego.

ppendix A. Supplementary data

Supplementary data associated with this article can be found, inhe online version, at doi:10.1016/j.ijmm.2010.02.004.

eferences

juebor, M.N., Flower, R.J., Hannon, R., Christie, M., Bowers, K., Verity, A., Perretti,M., 1998. Endogenous monocyte chemoattractant protein-1 recruits monocytesin the zymosan peritonitis model. J. Leukoc. Biol. 63, 108–116.

ltenhoefer, A., Oswald, S., Sonnenborn, U., Enders, C., Schulze, J., Hacker, J.,Oelschlaeger, T.A., 2004. The probiotic Escherichia coli strain Nissle 1917interferes with invasion of human intestinal epithelial cells by differententeroinvasive bacterial pathogens. FEMS Immunol. Med. Microbiol. 40,223–229.

rrach, N., Zhao, M., Porwollik, S., Hoffman, R.M., McClelland, M., 2008. Salmonellapromoters preferentially activated inside tumors. Cancer Res. 68, 4827–4832.

uzzola, F.R., Barbagelata, M.S., Caccuri, R.L., Sordelli, D.O., 2006. Attenuation andpersistence of and ability to induce protective immunity to a Staphylococcusaureus aroA mutant in mice. Infect. Immun. 74, 3498–3506.

hamberlain, L.M., Strugnell, R., Dougan, G., Hormaeche, C.E., Demarco de Hor-maeche, R., 1993. Neisseria gonorrhoeae strain MS11 harbouring a mutation ingene aroA is attenuated and immunogenic. Microb. Pathog. 15, 51–63.

herepanov, P.P., Wackernagel, W., 1995. Gene disruption in Escherichia coli: TcRand KmR cassettes with the option of Flp-catalyzed excision of the antibiotic-resistance determinant. Gene 158, 9–14.

atsenko, K.A., Wanner, B.L., 2000. One-step inactivation of chromosomal genesin Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. U.S.A. 97,6640–6645.

ougan, G., Maskell, D., Pickard, D., Hormaeche, C., 1987. Isolation of stable aroAmutants of Salmonella typhi Ty2: properties and preliminary characterisation inmice. Mol. Gen. Genet. 207, 402–405.

orbes, N.S., Munn, L.L., Fukumura, D., Jain, R.K., 2003. Sparse initial entrapment ofsystemically injected Salmonella typhimurium leads to heterogeneous accumu-lation within tumors. Cancer Res. 63, 5188–5193.

rozdanov, L., Zahringer, U., Blum-Oehler, G., Brade, L., Henne, A., Knirel, Y.A.,Schombel, U., Schulze, J., Sonnenborn, U., Gottschalk, G., Hacker, J., Rietschel,E.T., Dobrindt, U., 2002. A single nucleotide exchange in the wzy gene is respon-sible for the semirough O6 lipopolysaccharide phenotype and serum sensitivityof Escherichia coli strain Nissle 1917. J. Bacteriol. 184, 5912–5925.

ale, T.L., Sansonetti, P.J., Schad, P.A., Austin, S., Formal, S.B., 1983. Characterization ofvirulence plasmids and plasmid-associated outer membrane proteins in Shigellaflexneri, Shigella sonnei, and Escherichia coli. Infect. Immun. 40, 340–350.

eldin, C.H., Rubin, K., Pietras, K., Ostman, A., 2004. High interstitial fluid pressure– an obstacle in cancer therapy. Nat. Rev. Cancer 4, 806–813.

oiseth, S.K., Stocker, B.A., 1981. Aromatic-dependent Salmonella typhimurium arenon-virulent and effective as live vaccines. Nature 291, 238–239.

vins, B.E., Welkos, S.L., Knudson, G.B., Little, S.F., 1990. Immunization against anthraxwith aromatic compound-dependent (Aro-) mutants of Bacillus anthracis andwith recombinant strains of Bacillus subtilis that produce anthrax protectiveantigen. Infect. Immun. 58, 303–308.

ain, R.K., 1994. Barriers to drug delivery in solid tumors. Sci. Am. 271, 58–65.ain, R.K., 1998. The next frontier of molecular medicine: delivery of therapeutics.

Nat. Med. 4, 655–657.asinskas, R.W., Forbes, N.S., 2007. Salmonella typhimurium lacking ribose chemore-

ceptors localize in tumor quiescence and induce apoptosis. Cancer Res. 67,3201–3209.

ienle, G., Kiene, H., 2003. Die Coley‘sche Fiebertherapie der Krebserkrankung –historischer Markstein oder heute noch Vorbild? Der Merkurstab 56, 355–364.

otloff, K.L., Noriega, F., Losonsky, G.A., Sztein, M.B., Wasserman, S.S., Nataro, J.P.,Levine, M.M., 1996. Safety, immunogenicity, and transmissibility in humansof CVD 1203, a live oral Shigella flexneri 2a vaccine candidate attenuated bydeletions in aroA and virG. Infect. Immun. 64, 4542–4548.

ee, C.H., Wu, C.L., Shiau, A.L., 2004. Endostatin gene therapy delivered by Salmonellacholeraesuis in murine tumor models. J. Gene Med. 6, 1382–1393.

ee, C.H., Wu, C.L., Shiau, A.L., 2005a. Systemic administration of attenuatedSalmonella choleraesuis carrying thrombospondin-1 gene leads to tumor-specifictransgene expression, delayed tumor growth and prolonged survival in themurine melanoma model. Cancer Gene Ther. 12, 175–184.

ical Microbiology 300 (2010) 449–456

Lee, C.H., Wu, C.L., Tai, Y.S., Shiau, A.L., 2005b. Systemic administration of attenuatedSalmonella choleraesuis in combination with cisplatin for cancer therapy. Mol.Ther. 11, 707–716.

Lee, C.H., Wu, C.L., Chen, S.H., Shiau, A.L., 2009. Humoral immune responses inhibitthe antitumor activities mediated by Salmonella enterica serovar Choleraesuis.J. Immunother. 32, 376–388.

Leschner, S., Westphal, K., Dietrich, N., Viegas, N., Jablonska, J., Lyszkiewicz, M.,Lienenklaus, S., Falk, W., Gekara, N., Loessner, H., Weiss, S., 2009. Tumor invasionof Salmonella enterica serovar Typhimurium is accompanied by strong hemor-rhage promoted by TNF-alpha. PLoS One 4, e6692.

Loeffler, M., Le’Negrate, G., Krajewska, M., Reed, J.C., 2007. Attenuated Salmonellaengineered to produce human cytokine LIGHT inhibit tumor growth. Proc. Natl.Acad. Sci. U.S.A. 104, 12879–12883.

Low, K.B., Ittensohn, M., Luo, X., Zheng, L.M., King, I., Pawelek, J.M., Bermudes, D.,2004. Construction of VNP20009: a novel, genetically stable antibiotic-sensitivestrain of tumor-targeting Salmonella for parenteral administration in humans.Methods Mol. Med. 90, 47–60.

Mantovani, A., Sica, A., Sozzani, S., Allavena, P., Vecchi, A., Locati, M., 2004. Thechemokine system in diverse forms of macrophage activation and polarization.Trends Immunol. 25, 677–686.

Mengesha, A., Dubois, L., Chiu, R.K., Paesmans, K., Wouters, B.G., Lambin, P., Theys,J., 2007. Potential and limitations of bacterial-mediated cancer therapy. Front.Biosci. 12, 3880–3891.

Pawelek, J.M., Low, K.B., Bermudes, D., 1997. Tumor-targeted Salmonella as a novelanticancer vector. Cancer Res. 57, 4537–4544.

Pawelek, J.M., Low, K.B., Bermudes, D., 2003. Bacteria as tumour-targeting vectors.Lancet Oncol. 4, 548–556.

Platt, J., Sodi, S., Kelley, M., Rockwell, S., Bermudes, D., Low, K.B., Pawelek, J., 2000.Antitumour effects of genetically engineered Salmonella in combination withradiation. Eur. J. Cancer 36, 2397–2402.

Priebe, G.P., Brinig, M.M., Hatano, K., Grout, M., Coleman, F.T., Pier, G.B., Goldberg,J.B., 2002. Construction and characterization of a live, attenuated aroA dele-tion mutant of Pseudomonas aeruginosa as a candidate intranasal vaccine. Infect.Immun. 70, 1507–1517.

Roberts, M., Maskell, D., Novotny, P., Dougan, G., 1990. Construction and character-ization in vivo of Bordetella pertussis aroA mutants. Infect. Immun. 58, 732–739.

Royo, J.L., Becker, P.D., Camacho, E.M., Cebolla, A., Link, C., Santero, E., Guzman, C.A.,2007. In vivo gene regulation in Salmonella spp. by a salicylate-dependent controlcircuit. Nat. Methods 4, 937–942.

Sansonetti, P.J., Phalipon, A., 1999. M cells as ports of entry for enteroinvasivepathogens: mechanisms of interaction, consequences for the disease process.Semin. Immunol. 11, 193–203.

Schlee, M., Wehkamp, J., Altenhoefer, A., Oelschlaeger, T.A., Stange, E.F., Fellermann,K., 2007. Induction of human beta-defensin 2 by the probiotic Escherichia coliNissle 1917 is mediated through flagellin. Infect. Immun. 75, 2399–2407.

Stecher, B., Hapfelmeier, S., Muller, C., Kremer, M., Stallmach, T., Hardt, W.D.,2004. Flagella and chemotaxis are required for efficient induction of Salmonellaenterica serovar Typhimurium colitis in streptomycin-pretreated mice. Infect.Immun. 72, 4138–4150.

Stritzker, J., Janda, J., Schoen, C., Taupp, M., Pilgrim, S., Gentschev, I., Schreier, P.,Geginat, G., Goebel, W., 2004. Growth, virulence, and immunogenicity of Listeriamonocytogenes aro mutants. Infect. Immun. 72, 5622–5629.

Stritzker, J., Weibel, S., Hill, P.J., Oelschlaeger, T.A., Goebel, W., Szalay, A.A., 2007.Tumor-specific colonization, tissue distribution, and gene induction by probioticEscherichia coli Nissle 1917 in live mice. Int. J. Med. Microbiol. 297, 151–162.

Toso, J.F., Gill, V.J., Hwu, P., Marincola, F.M., Restifo, N.P., Schwartzentruber, D.J.,Sherry, R.M., Topalian, S.L., Yang, J.C., Stock, F., Freezer, L.J., Morton, K.E., Seipp, C.,Haworth, L., Mavroukakis, S., White, D., MacDonald, S., Mao, J., Sznol, M., Rosen-berg, S.A., 2002. Phase I study of the intravenous administration of attenuatedSalmonella typhimurium to patients with metastatic melanoma. J. Clin. Oncol. 20,142–152.

Van Rooijen, N., Sanders, A., 1994. Liposome mediated depletion of macrophages:mechanism of action, preparation of liposomes and applications. J. Immunol.Methods 174, 83–93.

Wei, M.Q., Ellem, K.A., Dunn, P., West, M.J., Bai, C.X., Vogelstein, B., 2007. Facultativeor obligate anaerobic bacteria have the potential for multimodality therapy ofsolid tumours. Eur. J. Cancer 43, 490–496.

Weibel, S., Stritzker, J., Eck, M., Goebel, W., Szalay, A.A., 2008. Colonization of exper-imental murine breast tumours by Escherichia coli K-12 significantly alters thetumour microenvironment. Cell. Microbiol. 10, 1235–1248.

Westphal, K., Leschner, S., Jablonska, J., Loessner, H., Weiss, S., 2008. Containment oftumor-colonizing bacteria by host neutrophils. Cancer Res. 68, 2952–2960.

Zhang, L., Gao, L., Zhao, L., Guo, B., Ji, K., Tian, Y., Wang, J., Yu, H., Hu, J., Kalvakolanu,D.V., Kopecko, D.J., Zhao, X., Xu, D.Q., 2007. Intratumoral delivery and sup-pression of prostate tumor growth by attenuated Salmonella enterica serovar

Zhao, M., Geller, J., Ma, H., Yang, M., Penman, S., Hoffman, R.M., 2007. Monother-apy with a tumor-targeting mutant of Salmonella typhimurium cures orthotopicmetastatic mouse models of human prostate cancer. Proc. Natl. Acad. Sci. U.S.A.104, 10170–10174.