The ‘biodrug’ concept: an innovative approach to therapy

TRENDS in Biotechnology Vol.19 No.10 October 2001

http://tibtech.trends.com 0167-7799/01/$ – see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S0167-7799(01)01739-5

393Review

Stéphanie Blanquet,

Eric Beyssac,

Michel Renaud

and Monique Alric*

Equipe de RechercheTechnologiqueConception, Ingénierie etDéveloppement del’Aliment et duMédicament, Faculté dePharmacie, 28 placeHenri-Dunant, 63000Clermont-Ferrand, France.*e-mail: [email protected]

Denis Pompon

Centre de GénétiqueMoléculaire, CNRS,Avenue de la Terrasse,91190 Gif/Yvette, France.

Sylvie Marol-Bonnin

Digestar company, BiopoleClermont-Limagne, 63360Saint-Beauzire, France.

The tools used for genetic engineering that have beendeveloped to date have led to the emergence of novelapplications using genetically modified microorganismsto produce drugs in large-scale bioprocesses1. Aninnovative extension of these approaches is drugproduction directly in the digestive environment byingested living recombinant microorganisms. For thispurpose, recombinant bacteria, mainly lactic acidbacteria, have been studied2,3. Here, we present yeastas a convenient host and a good alternative for theproduction of ‘BIODRUGS’ (see Glossary).

Yeasts have been widely used in food and beveragefermentation processes for centuries. The mostcommon yeasts, Saccharomyces cerevisiae andSaccharomyces boulardii, have a ‘GENERALLY

RECOGNIZED AS SAFE’ (GRAS) status and have recentlybeen used as PROBIOTICS in both animals and humans,and as BIOTHERAPEUTIC AGENTS in some humandigestive pathologies, such as antibiotic-associateddiarrhoea and Clostridium difficile-disease4. In thepast few decades, S. cerevisiae has become anattractive host for the production of recombinantproteins and bioconversion owing to its highproductivity and ease of genetic engineering.

The biodrug concept was validated5 using arecombinant model S. cerevisiae expressing the plantCYTOCHROME P450 73A1. This enzyme provides arelevant model of bioconversion for potentialtherapeutic applications, such as ‘biodetoxication’in the digestive environment. The yeasts have beenstudied in an artificial digestive system, whichsimulates human digestion. Here, we review thepotential uses of biodrugs in medical applications.

Genetically engineered microorganisms as a delivery

vector to the gastrointestinal tract: bacteria or yeasts?

The biodrug concept comprises two main ideas:(1) living microorganisms can carry out either

bioconversion or biosynthesis in the digestiveenvironment and (2) bioconversion can lead eitherto the production of a bioactive product or to theremoval of undesirable compounds. The bioactiveproducts resulting from bioconversion orbiosynthesis can be secreted in the digestivemedium, be bound to the cells or accumulate insidethe cells and be released in the digestive medium by cell lysis.

To use microorganisms as delivery vectors to thegastrointestinal tract, heterologous gene expressionstrategies that have already been developed have tobe adapted to particular constraints of the digestivetract environment. Promoters with an activity thatis adjustable and adapted to the digestiveenvironment have to be developed. One promisingapproach is the identification of suitable promotersusing DNA chips for global transcriptome analysisfollowing the culture of microorganism cells insimulated digestive environments. The resultinggenetically engineered cells would ensure acontrolled flow of the drug directly to its absorptionor reaction site. At the bioethics level, there is aneed for development of new ‘clean’ geneticengineering procedures allowing construction ofrecombinant microorganisms that are compatible

Cell engineering technology using recombinant microorganisms has created

new opportunities in the development of innovative drugs. This article

presents the use of living genetically engineered microorganisms, such as

bacteria or yeasts, as a new delivery vehicle to the gastrointestinal tract. This

‘biodrug’ concept was demonstrated using recombinant Saccharomyces

cerevisiae expressing the plant cytochrome P450 73A1. This enzyme provides a

relevant model for potential therapeutic applications, such as ‘biodetoxication’

in the digestive environment. An artificial gastrointestinal tract simulating

human digestion was chosen as a powerful tool to validate the biodrug

concept. This approach offers a novel strategy for drug discovery and testing.

The ‘biodrug’ concept: an innovative

approach to therapy

Stéphanie Blanquet, Sylvie Marol-Bonnin, Eric Beyssac, Denis Pompon, Michel Renaud

and Monique Alric

Biodrugs: bioactive drugs produced, at a level of thegastrointestinal tract, by living orally-administered recombinantmicroorganisms. These microorganisms can performbioconversion or biosynthesis in the digestive environment.Generally recognized as safe (GRAS): According to the US Foodand Drug Administration, a substance may be GRAS only if itsgeneral recognition of safety is based on the view of expertsqualified to evaluate the safety of the substance. GRAS status isbased either on a history of safe use in food before 1958 or onscientific procedures, which require the same quantity andquality of evidence as would be required to obtain a foodadditive regulation.Probiotics: living microorganisms belonging to the natural florawith low or no pathogenicity but with functions of importance tothe health and well being of the host.Biotherapeutic agents: referring to living microorganisms havingspecific therapeutic properties, for example for the treatment orprevention of infectious diseases.Cytochrome P450: the cytochrome P450 gene superfamilyencodes many isoenzymes that have major roles in thedetoxication system of the body by metabolising xenobiotics,such as drugs, procarcinogens (benzo(a)pyrene for example) and pesticides. They are also involved in the metabolism ofphysiologically important substances such as steroids,fat-soluble vitamins and fatty acids.

Glossary

with and risk-free in humans. The absence ofmobilizable vectors, antibiotic selection markersand bacterial sequences that could promote genetransfer to host bacteria is, therefore, essential. Inaddition, environmental confinement ofrecombinant cells could be achieved by introducinga suicide process6 (activation of a toxic protein orrepression of an essential gene) that triggers theelimination of the microorganisms upon leavingthe digestive tract.

Bacteria, as well as yeasts, can be considered asdelivery vectors to the gastrointestinal tract(Table 1). As unicellular microorganisms, both yeastsand bacteria offer several advantages with regard togenetic manipulation, culture facilities and low-costproduction. Also, numerous biotechnological toolsare available7. This is particularly the case forLactococcus lactis8, which has been sequenced,andfor S. cerevisiae9, the first eukaryotic microorganismto be fully sequenced. In the case of S. cerevisiae, forexample, promoters with various strengths andmodes of regulation have been characterized,numerous vectors have been identified and a widevariety of non-antibiotic-dependent selectionmarkers are available.

It is difficult to compare the advantages anddisadvantages of the different candidates forbiodrug delivery, other than the generally acceptedprinciples (Table 1), because many parameters must

be taken into account. These parameters include:(1) the microorganism; (2) the heterologous gene;(3) the genetic construction; and (4) the way todeliver the biodrug. However, in some situations,yeasts can be advantageous over bacteria for severalreasons. Although bacteria are mainly used toexpress simple heterologous proteins, yeasts areefficient alternative expression systems when aeukaryotic environment is required for thefunctional expression of a particular gene. This ismainly the case when membrane proteins ormulti-component complexes are involved, or whenpost-transcriptional processing (e.g. proteolysis,glycosylation) is necessary for the activity ortransport of the product. For example, yeast is awell-adapted host for the expression of multi-drugresistance (MDR) proteins and cytochrome P450enzymes10. In addition, co-expression of multipleheterologous proteins is possible in yeast cells,allowing the reconstruction of artificial metabolicpathways that can be coupled with natural orengineered endogenous biosynthetic pathways,leading to a self-sufficient strain for drugproduction11. Moreover, the absence in therecombinant yeasts of bacterial sequences that areliable to promote gene transfer to host bacteria canbe accomplished using PCR for cassette constructionand the efficient site-targeted homologousrecombination machinery of yeasts for cassette


http://tibtech.trends.com

394 Review

Table 1. Microorganisms as delivery vectors to the gastrointestinal tract: bacteria or yeasts?

Yeasts Bacteria

Safety 'GRAS' status for Saccharomyces cerevisiae, ‘GRAS' status for lactic acid bacteria such asSaccharomyces boulardii and Kluyveromyces Lactococci (L. lactis), Lactobacilli lactis. (Lb. acidophilus) and Bifidobacteria

Production Advantage: Easy to cultivate (unicellular Advantage: Easy to cultivate (unicellular microorganisms). microorganisms).

Genetic construction Advantages: Fully sequenced; S. cerevisiae9. Advantages: Fully sequenced; L. lactis8.Efficient site-targeted homologous recombination. Possibility of absence of mobilizable Possibility of absence of mobilizable vectors and vectors and antibiotic selection markers.antibiotic selection markers. Disadvantage: Bacterial sequences being

No bacterial sequences promoting gene transfer liable to promote gene transfer to hostto host bacteria. bacteria32.

Post-translational Advantages: Eukaryotic post-translational Disadvantages: No eukaryotic modifications modifications; required for eukaryotic proteins post-translational modifications (if they are

expression. required for activity of the protein).Disadvantage: Possibility of excessive or erroneous glycosylation3.

Physiology Advantage: Insensitive to most of the frequently Disadvantage: Sensitive to most of the used antibiotics (such as penicillin). frequently used antibiotics (such as

penicillin).

Survival rate in the High survival rate in in vitro stomach and small Low survival rate in in vitro stomach and gastro-intestinal tract intestine5. small intestine (see for various strains of

Disadvantages: Few data in humans34. lactic acid bacteria19).Advantage: Numerous data in human for Bifidobacteria and Lactobacilli35.

Way of Biodrugs delivery High survival rate and 'good condition' in the Low survival rate in the stomach and small stomach and small intestine: more suitable for intestine; more suitable for delivery bioconversion or secretion (see Ref. 5 for following cell lysis (see Ref. 13 for L. lactisS. cerevisiae in stomach and small intestine). in the duodenum).

integration into the genome12. Finally, yeasts are notsensitive to antibacterial products such as penicillin,allowing the simultaneous administration ofbiodrugs and antibiotic treatments. Yeast, especiallyS. cerevisiae, thus emerges as a promising candidatefor biodrug development.

The choice of the mode of biodrug delivery can be influenced by the survival rate of themicroorganisms in the gastrointestinal tract. For example, owing to their low survival rate in the upper part of the gastrointestinal tract, lacticacid bacteria, such as L. lactis13, are suitablecandidates for biodrug delivery following cell lysis.By contrast, owing to their high survival rate in thestomach and small intestine, yeasts, such asS. cerevisiae5, are suitable candidates forbioconversion or secretion, both requiringmicroorganisms to be in ‘good condition’.

The gastro-intestinal tract model: a powerful tool in

biodrug development

Several in vitro digestive models have beendeveloped but most of them have only a limitednumber of simulated parameters, arecompartmentally fragmented and lackdynamism14,15. A more complete artificial digestivesystem, composed of a stomach and small intestinemodel16 (Fig. 1a) and a large intestine model17

(Fig. 1b), was thus developed by the NederlandseOrganisatie voor Toegepast (TNO) Nutrition andFood Research Institute and validated incollaboration with the University of Auvergne(Auvergne, France). It is a multi-compartmental,dynamic, computer-controlled system that closelymimics the in vivo gastric and intestinal conditionsof humans and monogastric animals. The mainparameters of digestion, such as pH, temperature,peristaltic movements, gastric, biliary andpancreatic secretions, absorption of nutrients andwater, and microflora activity, are all reproduced asfaithfully as possible (Table 2). The model has beenvalidated by many studies including: (1) absorptionof the products of digestion16; (2) bioavailability ofminerals18; (3) survival rate of microorganisms19;and (4) composition and enzymatic activities of themicroflora, production and quantification of shortchain fatty acids and gases17. This system hasadvantages over research in human subjects orlaboratory animals with respect to accuracy, rapidity,continuous process control and data collection,modular set-up and high flexibility. In particular, it isan excellent alternative to animal experimentationin terms of reproducibility, easy manipulation andthe possibility of collecting samples at any momentduring digestion and at any level of the digestivetract, thereby increasing the amount of availabledata. In addition, potentially toxic molecules can betested with no ethical constraints.

Consequently, this in vitro model is a powerfulstudy tool in all stages of biodrug development

before clinical phases. The survival rate of themicroorganism cells and their physiological statecan be studied. In addition, the efficiency of newlydeveloped molecular tools (e.g. promoters, selectionmarkers and vectors) can be evaluated to optimizethe functionality of recombinant microorganisms inthe digestive environment. The in vitro model canalso aid the selection of pharmaceuticalformulations ensuring both the release ofmicroorganisms directly at their reaction site andtheir optimal activity (by addition of appropriatesubstrate and inductor). Experiments in theartificial digestive system will provide necessarydata on the biological safety of genetically modified microorganisms. For example, thepotential gene transfer to the human flora can be studied in the large intestine model(Havenaar, R., pers. commun.) and the cell deathoutside of the digestive tract can be checked toensure there is no dissemination in theenvironment. This system was therefore chosen tostudy the behaviour of genetically engineeredyeasts in the digestive environment in the firststages of the biodrug development.

Validation of the biodrug concept

As a model to validate the biodrug concept in termsof bioconversion5, a biological system involvingplant cytochrome P450 73A1 (cinnamate4-hydroxylase) and over-expressed yeast NADPHcytochrome P450 reductase (CPR) was chosen(Fig. 2a, 2b). The recombinant S. cerevisiae20

catalyses the bioconversion of trans-cinnamic acidinto p-coumaric acid, reproducing the first oxidativestep of the plant phenylpropanoid pathway.Genetically engineered yeasts and trans-cinnamicacid were introduced simultaneously into eitherthe stomach or the large intestine of thegastrointestinal tract model. The yeasts showed ahigh level of resistance to gastric and intestinalsecretions (Fig. 3a), but seemed to be more sensitiveto the conditions of the large intestine (Fig. 3b),probably owing to microflora competition. Thesurvival rate of the yeasts observed in the upperpart of the gastrointestinal tract model was muchhigher than that of lactic acid bacteria in closedexperimental conditions19. For the first time, theability of recombinant yeasts to performbioconversion was demonstrated in a simulatedhuman digestive environment: (1) in the stomachand small intestine model, 41 ± 6% (n=3) of thetrans-cinnamic acid ingested was converted intop-coumaric acid after 4 hours digestion (Fig. 3c),and (2) the bioconversion activity was also observedin the colon, but was weaker owing to parallelmetabolic pathways and cell death (unpublishedresults). The bioconversion reaction, which requiresmembrane enzymes, as well as NADPH and NADHco-factors (Fig. 2a), can occur only with yeast cellsthat are in ‘good condition’. This shows that this



395Review

method is efficient for biodrug release throughoutthe length of the digestive tract.

Previously, several studies involving bacteria, notyeasts, were performed in an animal or a human(only in the case of oral vaccine) digestiveenvironment (Table 3). For example, Prakash et al.21

used a rat model with renal failure to show thatorally-administered recombinant Escherichia coli

DH5 expressing Klebsiella aerogenes urease candecrease urea levels. Results showed that plasmauric acid is lowered from 71.00 ± 27.49 mg dl−1to20.33 ± 17.92 mg dl−1. In this study, recombinantE. coli were protected by alginate-poly-L-lysine-alginate microcapsules. Corthier et al. (pers.commun.) also demonstrated that recombinantL. lactis-expressing lipase could increase lipid



396 Review

TRENDS in Biotechnology

(a)

(b)

Duodenum

37°C

37°C

37°C

JejunumSodiumbicarbonate

Dialysates

Dialysates

Ileal delivery

Ilealabsorption

Hollowfibres

Jejunalabsorption

Electrolytes

Electrolytes

Ileum

pH electrode

Gastric secretions

Peristaltic valves

Peristalticvalves

StomachMeal

Flexible wall

Flexible wall

Hydrochloricacid/water

Billiary secretionsPancreatic secretions

Pression sensor

Pump

pH electrode

Pump

Sodium bicarbonate/electrolytes

Sodiumbicarbonate

Water

Water

Nitrogen under pressure/vacuum

Nitrogen

Nitrogen

Dialysis fibres

Faecal microflora

Nutritive medium

Level sensorSodium hydroxide

37°C

Fig. 1. A multi-compartmental, dynamic,computer-controlledgastrointestinal tractmodel that closely mimicsthe in vivo digestiveconditions of humanswas developed byNederlandse Organisatievoor Toegepast (TNO)Nutrition and FoodResearch Institute (Zeist,The Netherlands) andvalidated in collaborationwith the University ofAuvergne (Auvergne,France). It consists of (a) a stomach and smallintestine model16 and (b) alarge intestine model17

now constituted of threesimilar compartmentsthat are being validated toreproduce the three partsof the human colon(proximal, transverse anddistal). To simplify, onlyone of them is schematisedon the left of (b).

digestibility, following oral administration to pigwith pancreatic failure. In a murine digestiveenvironment, Fahl et al.22 demonstrated thebioconversion of the procarcinogenic benzo(a)pyrene[B(a)P] by E. coli co-expressing cytochrome P4501A1 and glutathione-S-transferase pi. They showedthat, after incubation of a meal containing B(a)Pand the recombinant E. coli in the stomach andduodenum of an anesthetized mouse, the mutagenicpotential of B(a)P was decreased, according tomutagenesis assay (Ames assay).

The ability of recombinant microorganisms tocarry out biosynthesis in the digestive tract has onlybeen demonstrated indirectly. Numerous studieshave shown that orally administered geneticallyengineered bacteria could induce an immuneresponse by delivering antigens directly in the

digestive tract. For example, live attenuatedrecombinant Salmonella have been studied forhuman vaccination against pathogens23 as reviewedby Bumann and colleagues24. Oral vaccines based onrecombinant Lactobacilli (e.g Lactobacillus casei25 todevelop an oral vaccine against anthrax) andLactococci (e.g. L. lactis26 to develop an oral vaccineagainst tetanus) have also been described. Moreover,Steidler et al.27 have recently shown a reduction ofcolitis symptoms in mice by the oral administrationof L. lactis secreting interleukin-10. To date, no workhas been published on the direct quantification ofproteins produced by living recombinantmicroorganisms in the digestive environment.A convenient model of biosynthesis is being studied inthe gastrointestinal tract system to further validatethis new opportunity.



397Review

Table 2. The gastrointestinal tract model

Physiological parameters Simulation Relevance and/or limits of the model

simulated by the model

Temperature The compartments are surrounded by water at 37°C.

pH The dynamic gastric pH curve is reproduced by secreting Relevance: The pH values are based on in vivo data, hydrochloric acid. monitored by computer and checked by pH electrodes in

The pH is kept to 6.5, 6.8 and 7.2 in duodenal, jejunal and each compartment. ileal compartments respectively, by secreting sodium Limit: The physiological evolution of pH in the colon bicarbonate. (increase of pH from proximal to distal colon), not yet

The pH is kept at 6 in the colon by secreting sodium reproduced, is being validated.hydroxide.

Peristaltic movements Peristaltic movements are mimicked by variation in water Relevance: These movements allow the mixing of chyme.pressure which is pumped into the space between the Limit: These movements are constant (in term of intensityglass jacket and the flexible wall (containing the chyme). and frequency).These changes are achieved by computer-activated rotary pumps.

Dynamic of chyme transit A mathematical model using power exponential equations36 Relevance: The dynamic of the physiological transit of is used to reproduce gastric and ileal deliveries. chyme is simulated. The parameters are based on in vivo

The transit of chyme is achieved by regulating the opening data (volume and composition of the meal, gastric and and closing sequences of peristaltic valve-pumps which ileal delivery, transit time).connect the compartments.

Volume of each compartment The initial volume in the stomach is ~300 ml. Limit: The length and volume of each compartment are The volume is maintained by pressure sensors at 30, 70 not the physiological ones.and 70 ml in the duodenum, jejunum and ileum Relevance: This limit is corrected by the residence time respectively. in each compartment.

The volume in each compartment of the colon is ~100 ml.

Gastric, biliary and pancreatic Pumps deliver pepsin and lipase in the stomach, Relevance: Secretion rates are based on in vivo data andsecretions pancreatine and biliary salts in the duodenum. regulated by computer-controlled pumps.

Limit: The enzymatic secretions of brush border cells (peptidases) are not reproduced.

Absorption of nutrients Semi-permeable membrane units (hollow fibres) are Relevance: The removal of digestion and fermentation and water connected to jejunal, ileal and colon compartments to products prevents inhibition of digestive enzymes and

absorb the products of digestion and fermentation, as microorganisms' activities respectively.well as water. Limit: The passive transport is simulated but not the

active one.

Microflora activity The colon is inoculated with a human faecal flora. Relevance: The stability of the microflora has been A nutritive medium, simulating ileal effluent medium, is proved. Validation of the colon17 was done with regards regularly introduced into the colon. to (a) the composition and enzymatic activities of the

microflora, (b) the concentration and composition of short chain fatty acids and (c) the gases production.

Oxygenation conditions The colon medium is kept under strictly anaerobic Limit: The physiological variations in oxygenation conditions with a flow of nitrogen. conditions from stomach to small intestine are not yet

reproduced.

Biodrug: from concept to medical applications

The biodrug concept involves the use of orallyadministered recombinant microorganisms as a newdrug delivery route to prevent or treat diseases. Thepotential applications are numerous and some ofthem have already been considered for bacteria(Table 3). However, yeasts can advantageouslyreplace bacteria for the reasons stated above(Table 1), particularly when eukaryotic organisationis required for the functional expression of aparticular gene.

Bioconversion

Three main types of potential applications can beconsidered. First, recombinant microorganismscould be administered orally to increase the body’sprotection against environmental xenobiotics,

particularly those borne by food (e.g. pesticides andchemical additives)22. Recently, widespread diseaseshave been linked to anomalies in humandetoxification processes. For instance, a deficiency of



398 Review


(a)

(b)

p GAL10-CYC1 CPR terYeast CPR ORF

URA3 p GAL10-CYC1

ADE2

YeDP60 plasmid

PGK ter

AmpRYeast 2µ ori E. coli ori

Plant CYP73A1 ORF

Coumarate

NADH NAD+

NADPH NADP+

CytosolCyt. b5

NADPH-Cyt.P450reductase

P450 73A1

CinnamateNADH-Cyt. b5

reductase

ER membrane

Engineered yeast CPR locus on genome of W(R) strain

Fig. 2. To validate the biodrug concept in terms of bioconversion, a recombinant modelSaccharomyces cerevisiae co-expressing the plant cytochrome P450 73A1 and the NADPHcytochrome P450 reductase (CPR) was used20. (a) A schematic representation of the multi-enzymatic complex. The cytochrome P450 73A1 and its associated proteins (CPR, endogenouscytochrome b5 and NADH cytochrome b5 reductase), required for electrons transfer, are bound tothe membrane of the endoplasmic reticulum (ER membrane). (b) The genetic construction of therecombinant model S. cerevisiae. The CYP73A1 open reading frame (ORF) from Jerusalemartichoke (Helianthus tuberosus) was placed under the transcriptional control of the galactoseinducible promoter GAL10-CYC1 (p GAL10-CYC1) and phosphoglycerate-kinase terminator(PGK ter). This expression cassette associated to the URA3 (orotate decarboxylase) and ADE2(amino imidazole ribonucleotide carboxylase) genes as complementation selection markers, to ayeast 2µ minicircle origin of replication (yeast 2µ ori) and to an E. coli replicon (Escherichia coli oriand AmpR as selection marker) constituted the expression vector YeDP60. This plasmid was used to transform S. cerevisiae W(R) strain (MAT α, Leu2, His3, Trp1, Ura3, Ade2, CanR, Cyr+), overproducingyeast CPR owing to substitution in the yeast genome of the natural CPR gene promoter by theGAL10-CYC1 inducible promoter.


(b)

0

20

40

60

80

100

120

0 1 2 3 4 5 6 7 8 9 10 11 12 13Time after introduction in the

large intestine (hour)

Via

ble

yeas

t cel

ls(%

of i

ntak

e)

(a)

0102030405060708090

100

0 60 120 180 240 300Time after introduction in the stomach (min)

Cum

ulat

ive

ileal

del

iver

y(%

of i

ntak

e)

(c)

0

10

20

30

40

50

0 60 120 180 240 300

Time after introduction in the stomach (min)

Tran

s-ci

nnam

ic a

cid

conv

ersi

on (

%)

Fig. 3. The survival rate of recombinant model Saccharomycescerevisiae and its ability to convert trans-cinnamic acid into p-coumaricacid have been studied in the gastro-intestinal tract model5.Trans-cinnamic acid and yeasts were introduced simultaneously ineither the stomach or the large intestine of the gastro-intestinal tractmodel. Recombinant yeasts were identified throughout the length ofthe digestive tract using PCR. Scoring of viable recombinant cells inthe digestive compartments was performed by enumeration on Petridishes following culture on selective medium. The bioconversionreaction was followed using a high performance liquidchromatography (HPLC) assay of trans-cinnamic and p-coumaric acidscontained in digestive samples. (a) A plot of the cumulative delivery ofliving recombinant yeasts from the ileal compartment compared withthe one of a nonabsorbable marker (blue circles, recombinant yeasts;green triangles, nonabsorbable marker). Values are expressed asmean percentages ± SD (n=3) of living yeasts relative to the amountintroduced in the stomach. (b) The survival rate of the recombinantyeasts in the large intestine. Values are expressed as mean percentages± SD (n=3) of living yeasts relative to the amount introduced in the largeintestine. (c) Bioconversion of trans-cinnamic acid into p-coumaric acidin the upper part of the gastro-intestinal tract model. The data obtainedat a given time in the different compartments (from stomach to ileum)were pooled to quantify the oxidative reaction during digestion.Values are expressed as mean percentages ± SD (n=3) of ingestedtrans-cinnamic acid converted into p-coumaric acid.

glutathione-S-transferase M1 (a detoxificationenzyme) has been associated with an increasedsusceptibility to various cancers, endometriosis andchronic bronchitis28. Therefore, by supplyingdeficient enzymes, recombinant microorganismscould be used to prevent or treat these multifactorialdiseases. The recombinant cells could also createsupplementary detoxification pathways, forexample by expressing anti-oxidative stressproteins, such as glutathione reductase orglutathione peroxidase. Second, biodrug mightconstitute an alternative to current therapy fordiseases that involves diffusing metabolites. Inparticular, the recombinant microorganisms couldcorrect errors of metabolism that result from gastricor intestinal enzyme deficiencies (e.g. lipase, trypsinand lactase) or organ failure (e.g. removing urea incase of kidney failure or ammonia in case of liverfailure2). The efficiency of microorganismsexpressing heterologous digestive enzymes can beeasily tested in the gastrointestinal tract model.Third, the recombinant cells could be used to controlthe activation of prodrug to drug, directly in thedigestive tract. This could be useful when the drug,but not the pro-drug, is either toxic at highconcentrations or damaged by digestive secretions.

Biosynthesis

Various therapeutic proteins, such as insulin,interleukins27, growth factors and coagulation

factors, could be produced by the geneticallymodified microorganisms in the digestive tract. Theadvantages of such a delivery system, comparedwith the classic systems, should be the followingpossibilities: (1) to administer drugs sensitive todigestive secretions; (2) to target specific sitesthroughout the length of the digestive tract; and(3) to provide therapeutic effects at lower doses.Another potential application is the development oforal vaccines29. These are based on the use ofrecombinant microorganisms delivering antigensdirectly into the digestive tract to stimulate a localimmune response (production of immunoglobulins)and prevent diseases. They could be used for vaccinesagainst bacterial pathogens23–25, bacterial toxins26,viruses and parasites or to control food allergies(e.g. bovine beta-lactoglobulin, a major cow’s milkallergen30). For example, yeasts expressing hepatitisB surface antigens have been suggested as a possibleoral vaccine against the virus31.

Conclusions

Using genetically engineered microorganisms, suchas bacteria or yeasts, as new delivery vehicles to thegastrointestinal tract is an important challenge forthe development of innovative drugs.

The potential medical applications of this newgeneration of biodrugs are numerous: for example,the correction of enzyme deficiencies, the control of theactivation of pro-drug to drug or the production of



399Review

Table 3. Biodrugs: from concept to medical applications

Biodrug concept Applications Examples Validation

Recombinant Experimental models Effect

microorganisms

BIOCONVERSION Biodetoxication Removal of E. coli (cytochrome P450 1A1 Stomach and duodenum Decrease of the mutagenicBenzo(a)pyrene and glutathione- of a mouse combined with potential of B(a)P.[(B(a)P)]22. S-transferase), in situ mutagenesis assay

administration. (Ames assay).

Model of S. cerevisiae (cytochrome Gastrointestinal tract Bioconversion of trans-biodetoxication5. P450 73A1), oral model simulating human cinnamic acid into

administration. digestion16,17. p-coumaric acid in all digestive compartments.

Correction of Correction of lipase L. lactis (lipase), oral Pig with pancreatic Increase of lipid errors of deficiency (Corthier, administration. deficiency. digestibility.metabolism G. et al., pers.

communication).

Correction of urease E. coli DH5 (urease), oral Rat with renal failure. Decrease of plasma uric deficiency21. administration (micro acid.

encapsulated).

BIOSYNTHESIS Synthesis of Secretion of L. lactis (interleukin-10), Mouse treated with Reduction of colitis biological interleukine-10 intragastric administration. dextran sulfate sodium symptoms.mediator (Ref. 27). (induction of chronic

colitis).Synthesis of oral Vaccination against L. lactis (tetanus toxin C57 BL/6 mouse. Induction of an immune

vaccine tetanus26. fragment C or TTFC), oral response such as serum administration. anti-TTFC antibody.

Vaccination against Attenuated Salmonella oral Mouse and rabbit. Induction of an immune Streptococcus (pneumococcal surface response such as serum pneumoniae23. protein A or PspA), oral anti-PspA antibody.

administration.

therapeutic proteins, such as vaccines, directly in thedigestive tract. In particular, by increasing the body’sprotection against environmental xenobiotics, thesebiodrugs can offer an innovative way to prevent or treatdiseases that escape traditional drug action, such ascancer or other widespread multifactorial diseases.

Although many microorganisms are potentialhosts for the production of biodrugs, the yeastemerges as a convenient candidate mainly becauseof its eukaryotic status and high level of resistanceto gastric and pancreatic secretions.

Concurrently, a gastrointestinal tract model16,17

that closely mimics the in vivo digestive conditions ofhumans has been developed and validated. Thismulti-compartmental, dynamic, computer-controlledsystem offers new opportunities for drug developmentand testing and will constitute a powerful alternativeto animal experimentation during all the pre-clinicalphases of biodrug development.

Consequently, for the first time, recombinantyeasts and the gastrointestinal tract model havebeen combined5 to provide a new method forassessing the feasibility of the biodrug concept.A bioconversion model was used involving the plantcytochrome P450 73A1, which catalyses thebioconversion of trans-cinnamic acid into p-coumaricacid. Scoring of viable recombinant cells wasperformed and their ability to oxidize trans-cinnamicacid was demonstrated in the different segments ofthe gastrointestinal tract model, thereby showingthe efficiency of this method for biodrug releasethroughout the length of the digestive tract. Theability of the yeast to carry out biosynthesis in situ isalso under consideration.

Choices of suitable microorganisms strains andcandidate genes, and the best way to administer theliving cells will be made according to the pathologicaltarget (bioconversion or biosynthesis).



400 Review

Acknowledgements

This work has beenselected by the FrenchMinistère de l’EducationNationale, de laRecherche et de laTechnologie in ‘appeld’offres post-génome1999’ and is supported bythis Ministry through aPhD grant. We particularlythank S. Rougeol for hertechnical assistance. Wethank also A. Meiniel,S. Rabot, R. Antonelli,J. Cohade and L. Laforêtfor their valuablecomments.

References

1 Primrose, S.B. (1986) The application ofgenetically engineered microorganisms in theproduction of drugs. J. Appl. Bacteriol.61, 99–116

2 Chang, T.M. and Prakash, S. (1998)Therapeutic uses of microencapsulatedgenetically engineered cells. Mol. Med. Today4, 221–227

3 Corthier, G. and Renault, P. (1999) Futuredirections for research on biotherapeutic agents:contribution of genetics approaches on lactic acidbacteria. In Biotherapeutic agents and infectionsdiseases (Elmer, GW., ed.), pp. 269–304, HumanaPress Inc., Totowa, NJ, USA

4 Mc Farland, L.V. and Bernasconi, P. (1993)Saccharomyces boulardii: a review of aninnovative biotherapeutic agent. Microb. Ecol.Health Dis. 6, 157–171

5 Alric, M. et al. (2000) Micro-organismes actifsdans l’environnement digestif, French patentn°0007843

6 Molin, S. (1993) Environmental potential of suicide genes. Curr. Opin. Biotechnol.4, 299–305

7 Kuipers, O.P. et al. (1997) Controlledoverproduction of proteins by lactic acidbacteria. Trends Biotechnol. 15, 135–140

8 Bolotin, A. et al. (2001) The complete genomesequence of the lactic acid bacteriumLactococcus lactis ssp. lactis IL1403.Genome Res. 11, 731–753

9 Goffeau, A. et al. (1996) Life with 6000 genes.Science 274, 563–567

10 Pompon, D. et al. (1997) Simulation of humanxenobiotic metabolism in micro-organisms.Yeast a good compromise between E. coli andhuman cells. J. Hepatol. 26, 81–85

11 Duport, C. et al. (1998) Self-sufficientbiosynthesis of pregnenolone and progesteronein engineered yeast. Nat. Biotechnol.16, 186–189

12 Tuite, M.F. (1992) Strategies for the geneticmanipulation of Saccharomyces cerevisiae.Crit. Rev. Biotechnol. 12, 157–188

13 Drouault, S. et al. (1999) Survival, physiology,and lysis of Lactococcus lactis in the digestivetract. Appl. Env. Microbiol. 65, 4881–4886

14 Savoie, L. (1994) Digestion and absorption offood: usefulness and limitations of in vitromodels. Can. J. Physiol. Pharmacol. 72, 407–414

15 Boisen, S. and Eggum, B.O. (1991) Criticalevaluation of in vitro methods for estimatingdigestibility in simple-stomach animals. Nutr.Res. Rev. 4, 141–162

16 Minekus, M. et al. (1995) A multicompartmentaldynamic computer-controlled model stimulatingthe stomach and small intestine. Atla 23,197–209

17 Minekus, M. et al. (1999) A computer-controlledsystem to simulate conditions of the largeintestine with peristaltic mixing, waterabsorption and absorption of fermentationproducts. Appl. Microbiol. Biotechnol.53, 108–114

18 Larsson, M. et al. (1997) Estimation of thebioavailability of iron and phosphorus in cerealsusing a dynamic in vitro gastrointestinal model.J. Sci. Food. Agric. 73, 99–106

19 Marteau, P. et al. (1997) Survival of lactic acidbacteria in a dynamic model of the stomach andsmall intestine: validation and the effects of bile.J. Dairy Sci. 80, 1031–1037

20 Urban, P. et al. (1994) Characterization ofrecombinant plant cinnamate 4-hydroxylaseproduced in yeast. Kinetic and spectralproperties of the major plant P450 of thephenylpropanoid pathway. Eur. J. Biochem.222, 843–850

21 Prakash, S. and Chang, T.M. (2000) In vitro andin vivo uric acid lowering by artificial cellscontaining microencapsulated geneticallyengineered E. coli DH5 cells. Int. J. Artif. Organs23, 429–435

22 Fahl, W.E. et al. (1999) Chemoprotectivebacterial strains, International patent n°WO 99/27953

23 Nayak, A.R. et al. (1998) A live recombinantavirulent oral Salmonella vaccine expressingpneumococcal surface protein A inducesprotective responses against Streptococcuspneumoniae. Infect. Immun. 66, 3744–3751

24 Bumann, D. et al. (2000) Recombinant liveSalmonella spp. for human vaccination againstpathogens. FEMS Immunol. Med. Microbiol.27, 357–364

25 Zegers, N.D. et al. (1999) Expression of theprotective antigen of Bacillus anthracis byLactobactillus casei: towards the developmentof an oral vaccine against anthrax. J. Appl. Microbiol. 87, 309–314

26 Robinson, K. et al. (1997) Oral vaccination ofmice against tetanus with recombinantLactococcus lactis. Nat. Biotechnol. 15, 653–657

27 Steidler, L. et al. (2000) Treatment of murinecolitis by Lactococcus lactis secretinginterleukin-10. Science 289, 1352–1355

28 Baranov, V.S. et al. (1996) Proportion of theGSTM1 0/0 genotype in some Slavicpopulations and its correlation with cysticfibrosis and some multifactorial diseases.Hum. Genet. 97, 516–520

29 Pouwels, P.H. et al. (1998) Lactic acid bacteriaas antigen delivery vehicles for oralimmunization purposes. Int. J. Food Microbiol.41, 155–167

30 Chatel, J.M. et al. (2001) Induction of mucosalimmune response after intranasal or oralinoculation of mice with Lactococcus lactisproducing bovine beta-lactoglobulin.Clin. Diagn. Lab. Immunol. 8, 545–551

31 Schreuder, M.P. et al. (1996) Yeast expressinghepatitis B virus surface antigen determinantson its surface: implications for a possible oralvaccine. Vaccine 14, 383–388

32 Gruzza, M. et al. (1994) Study of gene transferin vitro and in the digestive tract of gnotobioticmice from Lactococcus lactis strains to variousstrains belonging to human intestinal flora.Microb. Releases 2, 183–189

33 Tanner, W. and Lehle, L. (1987) Proteinglycosylation in yeast. Biochim. Biophys. Acta906, 81–99

34 Pecquet, S. et al. (1991) Kinetics ofSaccharomyces cerevisiae elimination from theintestines of Human volunteers and effect ofthis yeast on resistance to microbialcolonization in gnotobiotic mice. Appl. Env.Microbiol. 57, 3049–3051

35 Bezkorovainy, A. (2001) Probiotics:determinants of survival and growth in the gut.Am. J. Clin. Nutr. 73 (2 suppl), 399s–405s

36 Elashoff, J.D. et al. (1982) Analysis of gastricemptying data. Gastroenterology 83, 1306–1312

Documents

The ‘biodrug’ concept: an innovative approach to therapy