Plant-based Expression of Biopharmaceuticals

385

Plant-based Expression ofBiopharmaceuticals

J..org Kn

..ablein

Schering AG, Berlin, Germany

1 Introduction 387

2 Alternative Expression Systems 387

3 History of Plant Expression 389

4 Current Status of Plant-based Expression 3904.1 SWOT Analysis Reveals a Ripe Market for Plant Expression Systems 3904.2 Risk Assessment and Contingency Measures 392

5 The Way Forward: Moving Plants to Humanlike Glycosylation 396

6 Three Promising Examples: Tobacco (Rhizosecretion, Transfection)and Moss (Glycosylation) 398

6.1 Harnessing Tobacco Roots to Secrete Proteins 3986.2 High Protein Yields Utilizing Viral Transfection 3996.3 Simple Moss Performs Complex Glycosylation 401

7 Other Systems Used for Plant Expression 404

8 Analytical Characterization 405

9 Conclusion and Outlook 406

Acknowledgments 407

Bibliography 407Books and Reviews 407Primary References 407

Encyclopedia of Molecular Cell Biology and Molecular Medicine, 2nd Edition. Volume 10Edited by Robert A. Meyers.Copyright 2005 Wiley-VCH Verlag GmbH & Co. KGaA, WeinheimISBN: 3-527-30552-1

386 Plant-based Expression of Biopharmaceuticals

Keywords

GMPGood Manufacturing Practice (GMP) was established by WHO in 1968 to guaranteethe optimum degree of quality during production and processing of pharmaceuticals(cGMP means under the current regulations of the authorities).

TransgenicOrganisms that have externally introduced foreign DNA/genes stably integrated intotheir genome to, for example, produce desired substances like human insulin.

Plant-based ExpressionTransgenic plants can be genetically modified with a gene of interest to produce abiopharmaceutical of interest.

GlycosylationIt is the addition of polysaccharides to a certain molecule such as a protein. Themajority of proteins are synthesized in the rough endoplasmic reticulum (ER) wherethey undergo glycosylation.

BioreactorIt is a vessel in which a (bio)chemical process that involves organisms or biochemicallyactive substances (e.g. enzymes) derived from such organisms is carried out.

� Biopharmaceuticals are currently the mainstay products of the biotechnology marketand represent the fastest growing and, in many ways, the most exciting sectorwithin the pharmaceutical industry. The term ‘‘biopharmaceutical’’ was originatedin the 1980s, when a general consensus evolved that it represented a class oftherapeutics produced by means of modern biotechnologies. Already a quarter ofa century ago, ‘‘humulin’’ (recombinant human insulin, produced in E. coli anddeveloped by Genentech in collaboration with Eli Lilly) was approved and receivedmarketing authorization in the United States of America in 1982. Since then themarket for biopharmaceuticals has been steadily growing and currently nearly 150biopharmaceuticals have gained approval for general human use (EU and USA). Overthis period it became obvious that production capacities for biopharmaceuticals with‘‘conventional’’ bioreactors would be a bottleneck and that worldwide fermentationcapacities are limited. One exciting solution to these ‘‘capacity crunches’’ is theuse of transgenic plants to produce biopharmaceuticals. This article describesdifferent plant expression systems, their advantages and limitations, and concludesby considering some of the innovations and trends likely to influence the future ofplant-based biopharmaceuticals.

Plant-based Expression of Biopharmaceuticals 387

1Introduction

Biopharmaceuticals, which are large mole-cules produced by living cells, are currentlythe mainstay products of the biotechnologyindustry. Indeed, biologics such as Genen-tech’s (Vacaville, CA, USA) human growthfactor somatropin or Amgen’s (ThousandOaks, CA, USA) recombinant erythropoi-etin (EPO) have shown that biopharma-ceuticals can benefit a huge number ofpatients and also generate big profits forthese companies at the same time. Thesingle most lucrative product is EPO andcombined sales of the recombinant EPOproducts ‘‘Procrit’’ (Ortho biotech) and‘‘Epogen’’ (Amgen) have reportedly sur-passed the $6.5 billion mark. But it hasalso become obvious over the last coupleof years that current fermentation capaci-ties will not be sufficient to manufacture allbiopharmaceuticals (in the market alreadyor in development), because the marketand demand for biologics is continuouslyand very rapidly growing; for antibodiesalone (with at least 10 monoclonal an-tibodies approved and being marketed),the revenues are predicted to expand toUS$3 billion in 2002 and US$8 billion in2008. The 10 monoclonal antibodies onthe market consume more than 75% of theindustry’s manufacturing capability. Andthere are up to 60 more that are expectedto reach the market in the next six or sevenyears. Altogether, there are about 1200protein-based products in the pipeline witha 20% growth rate and the market forcurrent and late stage (Phase III) is es-timated to be US$42 billion in 2005 andeven US$100 billion in 2010. But, thereare obvious limitations of large-scale man-ufacturing resources and production ca-pacities – and pharmaceutical companies

are competing (see ref Knablein (2004),review).

To circumvent this capacity crunch, itis necessary to look into other technolo-gies rather than the established ones, like,for example, Escherichia coli or CHO (Chi-nese hamster ovary) cell expression. Onesolution to avoid these limitations couldbe the use of transgenic plants to ex-press recombinant proteins at low cost,in GMP (good manufacturing practice)quality greenhouses (with purification andfill finish in conventional facilities). Plantstherefore provide an economically soundsource of recombinant proteins, such as in-dustrial enzymes, and biopharmaceuticals.Furthermore, using the existing infrastruc-ture for crop cultivation, processing, andstorage will reduce the amount of capitalinvestment required for commercial pro-duction. For example, it was estimated thatthe production costs of recombinant pro-teins in plants could be between 10 and50 times lower than those for producingthe same protein in E. coli and Alan Dovedescribes a factor of thousand for cost ofprotein (US dollar per gram of raw mate-rial) expressed in, for example, CHO cellscompared to transgenic plants. So, at thedawn of this new millennium, a solution isimminent to circumvent expression capac-ity crunches and to supply mankind withthe medicines we need. Providing the rightamounts of biopharmaceuticals can nowbe achieved by applying our knowledge ofmodern life sciences to systems that wereon this planet long time before us – plants.

2Alternative Expression Systems

Currently, CHO cells are the most widelyused technology in biomanufacturing be-cause they are capable of expressingeukaryotic proteins (processing, folding,


and posttranslational modifications) thatcannot be provided by E. coli. A longtrack record exists for CHO cells, butunfortunately they bring some problemsalong when it comes to scaling up pro-duction. Transport of oxygen (and othergases) and nutrients is critical for the fer-mentation process, as well as the fact thatheat must diffuse evenly to all culturedcells. According to the Michaelis–Mentenequation, the growth rate depends on theoxygen/nutrient supply; therefore, goodmixing and aeration are a prerequisitefor the biomanufacturing process and areusually achieved by different fermentationmodes (see Fig. 1). But the laws of physicsset strict limits on the size of bioreac-tors. For example, an agitator achievesgood heat flow and aeration, but with in-creased fermenter size, shear forces alsoincrease and disrupt the cells – and build-ing parallel lines of bioreactors multipliesthe costs linearly. A 10 000-L bioreactorcosts between US$ 250 000 to 500 000and takes five years to build (concep-tual planning, engineering, construction,validation, etc.). An error in estimating de-mand for, or inaccurately predicting theapproval of, a new drug can be incrediblycostly. To compound the problem, regu-lators in the United States and Europedemand that drugs have to be producedfor the market in the same system used to

produce them for the final round of clinicaltrials, in order to guarantee bioequivalence(e.g. toxicity, bioavailability, pharmacoki-netics, and pharmacodynamics) of themolecule. So, companies have to choosebetween launching a product manufac-tured at a smaller development facility(and struggling to meet market demands)or building larger, dedicated facilities for adrug that might never be approved!

Therefore, alternative technologies areused for the expression of biopharma-ceuticals, some of them also at lowercosts involved (see Fig. 2). One such al-ternative is the creation of transgenicanimals (‘‘pharming’’), but this suffersfrom the disadvantage that it requires along time to establish such animals (ap-proximately 2 years). In addition to that,some of the human biopharmaceuticalscould be detrimental to the mammal’shealth, when expressed in the mammaryglands. This is why ethical debates some-times arise from the use of transgenicmammals for production of biopharma-ceuticals. Although there are no ethicalconcerns involved with plants, there aresocietal ones that will be addressed later.Another expression system (see Fig. 2) uti-lizes transgenic chicken. The eggs, fromwhich the proteins are harvested, arenatural protein-production systems. Butproduction of transgenic birds is still

(a)

Mechanic: Agitator Pneumatic: Gassing Hydrodynamic: Pumps

(b) (c) (d)

Fig. 1 Different fermentation modes forbioreactors. In order to achieve best aeration andmixing and to avoid high shear forces, differentfermentation modes are applied. (a) mechanical,(b) pneumatical, (c) hydrodynamic pumps,

(d) airlift reactor. Source: Kn..ablein J.

(2002) Transport Processes in Bioreactors andModern Fermentation Technologies, Lecture atUniversity of Applied Sciences,Emden, Germany.


Major technology

companies

$150 $1−$2$1−$2 $0.05Estimated cost(cost/g raw material)*

*Company estimates

Mammalian (CHO) cells

Amgen (Thousand Oaks, CA)Genentech (S. San Francisco, CA) other current biologics manufacturers:Crucell (Leiden, Netherlands) uses human cells

Transgenic mammal milk Transgenic chicken eggs Transgenic plants

Croptech (Blacksburg, VA)Epicyte (San Diego, CA)Large Scale Biology (Owensboro, KY)Meristem Therapeutics (Clermont-Ferrand, France)Prodigene (College Station, TX)

GTC Biotherapeutics (Framingham, MA)PPL Therapeutics (Edinburgh, UK)BioProtein (Paris, France)

Avigenics (Athens, GA)Origen Therapeutics (Burlingame, CA)TranXenoGen (shrewsbury, MA)Viragen (Plantation, FL)GeneWorks (Ann Arbor, MI)Vivalis (Nantes, France)

Fig. 2 Companies and technologies inbiomanufacturing. A comparison of differentexpression systems shows the big differences interms of costs, ranging from 150 US$ per gram

for CHO cells to 0.05 US$ per gram fortransgenic plants. Source: Dove, A.(2002) Uncorking the biomanufacturingbottleneck, Nat. Biotechnol. 20, 777–779.

several years behind transgenic mammaltechnology. Intensive animal housing con-straints also make them more susceptibleto disease (e.g. Asia 1997 or Europe 2003:killing of huge flocks with thousands ofchicken suffering from fowl pest). In thelight of development time, experience,costs, and ethical issues, plants are there-fore the favored technology, since suchsystems usually have short gene-to-proteintimes (weeks), some are already well es-tablished, and as mentioned before, theinvolved costs are comparatively low. Thislow cost of goods sold (COGS) for plant-derived proteins is mainly due to lowcapital costs: greenhouse costs are onlyUS$ 10 per m2 versus US$ 1000 per m2

for mammalian cells.

3History of Plant Expression

Plants have been a source of medicinalproducts throughout human evolution.These active pharmaceutical compounds

have been primarily small molecules,however. One of the most popular ex-amples is aspirin (acetylsalicylic acid) torelieve pain and reduce fever. A Frenchpharmacist first isolated natural salicin(a chemical relative of the compoundused to make aspirin) from white wil-low bark in 1829. Advances in geneticengineering are now allowing for theproduction of therapeutic proteins (as op-posed to small molecules) in plant tissues.Expression of recombinant proteins inplants has been well documented sincethe 1970s and has slowly gained credi-bility in the biotechnology industry andregulatory agencies. The first proof ofconcept has been the incorporation ofinsect and pest resistance into grains.For example, ‘‘Bt corn’’ contains genesfrom Bacillus thuringensis and is currentlybeing grown commercially. Genetic engi-neering techniques are now available forthe manipulation of almost all commer-cially valuable plants. Easy transforma-tion and cultivation make plants suitable


for production of virtually any recombi-nant protein.

Plants have a number of advantagesover microbial expression systems, butone of them is of outmost importance:they can produce eukaryotic proteins intheir native form, as they are capableof carrying out posttranslational modifi-cations required for the biological activityof many such proteins (see Fischer Schill-berg (2004), books). These modificationscan be acetylation, phosphorylation, andglycosylation, as well as others. Per se,there is no restriction to the kind of pro-teins that can be expressed in plants:vaccines (e.g. pertussis or tetanus toxins),serum proteins (e.g. albumin), growth fac-tors (e.g. vascular endothelial growth factor(VEGF), erythropoietin), or enzymes (e.g.urokinase, glucose oxidase, or glucocere-brosidase). However, enzymes sometimeshave very complex cofactors, which are es-sential for their catalytic mode of action,but cannot be supplied by most expressionsystems. This is why, for the expression ofsome enzymes, expression systems withspecial features and characteristics needto be developed. Another very importantclass of proteins is the antibodies (e.g.scFv, Fab, IgG, or IgA). More than 100antibodies are currently used in clinicaltrials as therapeutics, drug delivery vehi-cles, in diagnostics and imaging, and indrug discovery research for both screen-ing and validation of targets. Again, plantsare considered as the system of choicefor the production of antibodies (‘‘plan-tibodies’’) in bulk amounts at low costs.Since the initial demonstration that trans-genic tobacco (Nicotiana tabacum) is ableto produce functional IgG1 from mouse,full-length antibodies, hybrid antibodies,antibody fragments (Fab), and single-chainvariable fragments (scFv) have been ex-pressed in higher plants for a number of

purposes. These antibodies can serve inhealth care and medicinal applications, ei-ther directly by using the plant as a foodingredient or as a pharmaceutical or diag-nostic reagent after purification from theplant material. In addition, antibodies mayimprove plant performance, for example,by controlling plant disease or by modify-ing regulatory and metabolic pathways.

4Current Status of Plant-based Expression

4.1SWOT Analysis Reveals a Ripe Market forPlant Expression Systems

When I analyzed the different expres-sion systems regarding their strengths,weaknesses, opportunities, and threats(SWOT), the advantages of plants and theirpotential to circumvent the worldwide ca-pacity limitations for protein productionbecame quite obvious (see Fig. 3). Compar-ison of transgenic animals, mammaliancell culture, plant expression systems,yeast, and bacteria shows certain advan-tages for each of the systems. In theorder in which the systems were just men-tioned, we can compare them in termsof their development time (speed). Trans-genic animals have the longest cycle time(18 months to develop a goat), followedby mammalian cell culture, plants, yeast,and bacteria (one day to transform E.coli). If one looks at operating and capi-tal costs, safety, and scalability, the datashow that plants are beneficial: therefore,in the comparison (see Fig. 3), they areshown on the right-hand side already.But even for glycosylation, multimeric as-sembly and folding (where plants are notshown on the right-hand side, meaningother systems are advantageous), someplant expression systems are moving in


Speed

Minus

Operatingcost

Captialcosts

Glyco-sylation

Multimericassembly

Folding

Safety

Scalability

Bacteria YeastTRENDS in Biotechnology Vol.20 No.12, 2002Plants

Transgenicanimals

Mammaliancell culture

PlusStrengthsAccess new manufacturing facilities

High production rates/high protein yield

Relatively fast 'gene to protein' time

Safety benefits;no human pathogens/no TSE

Stable cell lines/high genetic stability

Simple medium (water, minerals & light)

Easy purification (ion exchange vs protA)

•

••

•

•

•

•

•

•

•

•

•

•

•

WeaknessesNo approved products yet (but Phase III)

No final guidelines yet (but drafts available)

Opportunities

Reduce projected COGS

Escape capacity limitations

Achieve human-like glycosylation

Threats

Food chain contamination

Segregation risk

Fig. 3 SWOT analysis of plant expressionsystems. Plant expression systems have a lot ofadvantages (plus) over other systems and aretherefore mostly shown on the right-hand side ofthe picture (Raskin, I., Fridlender, B., et al.(2002) Plants and human health in thetwenty-first century, Trends Biotechnol. 20,522–531). Herein different systems (transgenicanimals, mammalian cell culture, plants, yeast,and bacteria) are compared in terms of speed(how quickly they can be developed), operatingand capital costs, and so on, and plants areobviously advantageous. Even for glycosylation,

assembly, and folding, where plants are notshown on the right-hand side (meaning othersystems are advantageous), some plantexpression systems are moving in that direction(as will be shown exemplarily in the section onmoss). Also the weaknesses and threats can bedealt with, using the appropriate plantexpression system. Source: Kn

..ablein J. (2003)

Biotech: A New Era In The NewMillennium – From Plant Fermentation To PlantExpression Of Biopharmaceuticals, PDAInternational Congress, Prague, Czech Republic.

that direction. An example of this is themoss system from the company greenova-tion Biotech GmbH (Freiburg, Germany),which will be discussed in detail in theexample section. This system performsproper folding and assembly of even suchcomplex proteins like the homodimericVEGF. Even the sugar pattern could suc-cessfully be reengineered from plant tohumanlike glycosylation.

In addition to the potential of perform-ing human glycosylation, plants also enjoythe distinct advantage of not harboring

any pathogens, which are known to harmanimal cells (as opposed to animal cell cul-tures and products), nor do the productscontain any microbial toxins, TSE (Trans-missible Spongiform Encephalopathies),prions, or oncogenic sequences. In fact,humans are exposed to a large, con-stant dose of living plant viruses in thediet without any known effects/illnesses.Plant production of protein therapeuticsalso has advantages with regard to theirscale and speed of production. Plantscan be grown in ton quantities (using


existing plant/crop technology, like com-mercial greenhouses), be extracted withindustrial-scale equipment, and producekilogram-size yields from a single plot ofcultivation. These economies of scale areexpected to reduce the cost of productionof pure pharmaceutical-grade therapeuticsby more than 2 orders of magnitude ver-sus current bacterial fermentation or cellculture reactor systems (plus raw materialCOGS are estimated to be as low as 10%of conventional cell culture expenses).

Although a growing list of heterolo-gous proteins were successfully producedin a number of plant expression systemswith their manifold advantages, there arealso obvious downsides. One weakness isthat no product has been approved forthe market yet (but will be soon, sincesome are in Phase III clinical trials al-ready, see Table 1). The other weaknessis that no final regulatory guidelines ex-ist. But as mentioned before, regulatoryauthorities (Food and Drug Administra-tion (FDA), European Medicine EvaluationAgency (EMEA), and Biotechnology Regu-latory Service (BRS) and the BiotechnologyIndustry Organization (BIO) have draftedguidelines on plant-derived biopharma-ceuticals (see Table 2) and have askedthe community for comments. The FDAhas also issued several PTC (Points ToConsider) guidelines about plant-based bi-ologics, and review of the July 2002 PTCconfirms that the FDA supports this fieldand highlights the benefits of plant ex-pression systems – including the absenceof any pathogens to man from plant ex-tracts. The main concerns of using plantexpression systems are societal ones aboutenvironmental impacts, segregation risk,and contamination of the food chain.But these threats can be dealt with, us-ing nonedible plants (nonfood, nonfeed),

applying advanced containment technolo-gies (GMP greenhouses, bioreactors) andavoiding open-field production.

Owing to the obvious strengths of plantexpression systems, there has been explo-sive growth in the number of start-upcompanies. Since the 1990s, a numberof promising plant expression systemshave been developed, and in response tothis ‘‘blooming field’’ big pharmaceuticalcompanies have become more interested.Now, the plant expression field is ‘‘ripe’’for strategic alliances, and, in fact, thelast year has seen several major biotechcompanies begin partnerships with suchplant companies. The selection of severalsuch partnerships shown in Table 1 clearlydemonstrates that, in general, there hasbeen sufficient experimentation with var-ious crops to provide the overall proof ofconcept that transgenic plants can pro-duce biopharmaceuticals. However, andthis can be seen in the table as well, thecommercial production of biopharmaceu-ticals in transgenic plants is still in theearly stages of development and yet themost advanced products are in Phase IIIclinical development.

4.2Risk Assessment and ContingencyMeasures

For a number of reasons, including theknowledge base developed on geneticallymodifying its genome, industrial pro-cesses for extracting fractionated productsand the potential for large-scale produc-tion, the preferred plant expression systemhas been corn. However, the use ofcorn touches on a potential risk: someenvironmental activist groups and tradeassociations are concerned about the effecton the environment and possible contam-ination of the food supply. These issues


Tab.

1Pl

ant-

deri

ved

biop

harm

aceu

tical

sin

clin

ical

tria

ls.

Com

pany

Part

ner

Prot

ein/

indi

cati

onH

ost

Stag

e

Mon

sant

oG

uy’s

Hos

pita

lLon

don

Ant

icar

ies

antib

ody

Cor

nPh

ase

III

Larg

eSc

ale

Bio

logy

Ow

npr

oduc

tsc

Fv(n

on-H

odgk

in)

Toba

cco

Phas

eII

IsM

eris

tem

Ther

apeu

tics

Solv

ayPh

arm

aceu

tical

sG

astr

iclip

ase

Cor

nPh

ase

IILa

rge

Scal

eB

iolo

gyPr

odiG

ene,

Plan

tBio

scie

nce

Ant

i-ide

otyp

ean

tibod

yTo

bacc

oPh

ase

IM

onsa

nto

Neo

Rx

Ant

itum

oran

tibod

yC

orn

Phas

eI

Prod

iGen

eO

wn

prod

uct

TGEV

vacc

ine

Cor

nPh

ase

IEp

icyt

ePh

arm

aceu

tical

Dow

,Cen

toco

rA

nti-H

SVan

tibod

yC

orn

Phas

eI

Cro

pTe

chIm

mun

exEn

brel

(art

hriti

s)To

bacc

oPr

eclin

ical

Cro

pTe

chA

mge

nTh

erap

eutic

antib

odie

sTo

bacc

oPr

eclin

ical

Alta

Gen

Bio

scie

nce

Inc.

U.S

.Arm

y3

+bi

otec

hsA

ntib

odie

sPo

tato

Prec

linic

alM

eris

tem

Ther

apeu

tics

CN

RS

Hum

anla

ctof

erri

nC

orn

Prec

linic

alM

PBC

olog

neG

mbH

Ave

ntis

Cro

pSci

ence

Con

fiden

tial

Pota

toPr

eclin

ical

Tab.

2D

raft

edgu

idel

ines

onpl

ant-

deri

ved

biop

harm

aceu

tical

s.

Age

ncy

Gui

delin

eSt

atus

BR

S(B

iote

chno

logy

Reg

ulat

ory

Serv

ices

‘‘Cas

est

udy

onpl

ant-

deri

ved

biol

ogic

s’’f

orO

ffice

ofSc

ienc

ean

dTe

chno

logy

Polic

y/C

ounc

ilon

Envi

ronm

enta

lQua

lity

Rel

ease

d:M

ar5,

2001

BIO

(Bio

tech

nolo

gyIn

dust

ryO

rgan

izat

ion)

‘‘Ref

eren

ceD

ocum

entf

orC

onfin

emen

tan

dD

evel

opm

ento

fPla

nt-m

ade

Phar

mac

eutic

als

inth

eU

nite

dSt

ates

’’R

elea

sed:

May

17,2

002

BIO

(Bio

tech

nolo

gyIn

dust

ryO

rgan

izat

ion)

‘‘BIO

Posi

tion

onG

eogr

aphi

cR

estr

ictio

nsfo

rPl

ant-

mad

ePh

arm

aceu

tical

san

dIn

dust

rial

s’’

Rel

ease

d:O

ct22

,200

2

EMEA

(Eur

opea

nM

edic

ine

Eval

uatio

nA

genc

y)‘‘C

once

ptPa

per

onth

eD

evel

opm

ento

faC

omm

ittee

for

Prop

riet

ary

Med

icin

alPr

oduc

ts(C

PMP)

Poin

tsto

Con

side

ron

the

Use

ofTr

ansg

enic

Plan

tsin

the

Man

ufac

ture

ofB

iolo

gica

lMed

icin

alPr

oduc

tsfo

rH

uman

Use

’’

Rel

ease

d:M

ar01

,200

1

FDA

(Foo

dan

dD

rug

Adm

inis

trat

ion)

‘‘Dru

gs,B

iolo

gics

,and

Med

ical

Dev

ices

Der

ived

from

Bio

engi

neer

edPl

ants

for

Use

inH

uman

san

dA

nim

als’

’Is

sued

:Sep

6,20

02

EMEA

(Eur

opea

nM

edic

ine

Eval

uatio

nA

genc

y)‘‘P

oint

sTo

Con

side

rQ

ualit

yA

spec

tsof

Med

icin

alPr

oduc

tsco

ntai

ning

activ

esu

bsta

nces

prod

uced

byst

able

tran

sgen

eex

pres

sion

inhi

gher

plan

ts’’

Issu

ed:M

ar13

,200

2


are reflected in the regulatory guidelinesand have been the driving force to inves-tigate other plants as well. While manymature and larger companies have beenworking in this area for many years, thereare a number of newcomers that are de-veloping expertise as well. These smallercompanies are reacting to the concernsby looking at the use of nonedible plantsthat can be readily raised in greenhouses.All potential risks have to be assessedand contingency measures need to be es-tablished. Understanding the underlyingissues is mandatory to make sophisticateddecisions about the science and subse-quently on the development of appropriateplant expression systems for production ofbiopharmaceuticals.

Ongoing public fears from the foodindustry and the public, particularly inEurope (‘‘Franken Food’’) could havespillover effects on plant-derived pharma-ceuticals. Mistakes and misunderstand-ings have already cost the geneticallyenhanced grain industry hundreds of mil-lions of dollars. The only way to preventplant expression systems from suffer-ing the same dilemma is to provide thepublic with appropriate information onemerging discoveries and newly developedproduction systems for biopharmaceuti-cals. Real and theoretical risks involvethe spread of engineered genes into wildplants, animals, and bacteria (horizontaltransmission). For example, if herbicideresistance was transmitted to weeds, orantibiotic resistance was to be transmittedto bacteria, superpathogens could result. Ifthese genetic alterations were transmittedto their progeny (vertical transmission), anexplosion of the pathogens could cause ex-tensive harm. An example of this occurredseveral years ago, when it was feared thatpest-resistant genes had been transmit-ted from Bt corn to milkweed – leading

to the widespread death of Monarch but-terflies. Although this was eventually notfound to be the case, the public outcryover the incident was a wake-up call tothe possible dangers of transgenic foodtechnology. To avoid the same bad per-ception for biopharmaceuticals expressedin plants, there is the need for thor-ough risk assessment and contingencyplanning. One method is the employ-ment of all feasible safety strategies toprevent spreading of engineered DNA (ge-netic drift), like a basic containment ina greenhouse environment. Although nopractical shelter can totally eradicate insectand rodent intrusion, this type of isolationis very effective for self-pollinators andthose plants with small pollen dispersalpatterns. The use of species-specific, frag-ile, or poorly transmissible viral vectorsis another strategy. Tobacco mosaic virus(TMV), for example, usually only infects atobacco host.

It requires an injury of the plant togain entry and cause infection. Destructionof a field of TMV-transformed tobaccorequires only plowing under or applicationof a herbicide. These factors prevent bothhorizontal and vertical transmission. Inaddition, there is no known incidence ofplant viruses infecting animal or bacterialcells. Another approach is to avoid stabletransgenic germlines and therefore mostuses of transforming viruses do notinvolve the incorporation of genes intothe plant cell nucleus. By definition, it isalmost impossible for these genes to betransmitted vertically through pollen orseed. The engineered protein product isproduced only by the infected generationof plants. Another effective way to reducethe risk of genetic drift is the use ofplants that do not reproduce withouthuman aid. The modern corn plantcannot reproduce without cultivation and


the purposeful planting of its seeds.If a plant may sprout from grain, itstill needs to survive the wintering-overprocess and gain access to the properplanting depth. This extinction process isso rapid, however, that the errant loss ofan ear of corn is very unlikely to growa new plant. Another very well-knownexample of self-limited reproduction isthe modern banana. It propagates almostexclusively through vegetative cloning (i.e.via cuttings).

Pollination is the natural way for mostplants to spread their genetic information,make up new plants, and to deliver theiroffspring in other locations. The use ofplants with limited range of pollen dis-persal and limited contact with compatiblewild hosts therefore is also very effectiveto prevent genetic drift. Corn, for exam-ple, has pollen, which survives for only10 to 30 min and, hence, has an effectivefertilizing radius of less than 500 m. InNorth America, it has no wild-type rela-tives with which it could cross-pollinate.In addition to being spatially isolated fromnearby cornfields, transgenic corn can be‘‘temporally isolated’’ by being plantedat least 21 days earlier or 21 days laterthan the surrounding corn, to ensurethat the fields are not producing flowersat the same time. Under recent USDA(U.S. Department of Agriculture) regula-tions, the field must also be planted withequipment dedicated to the geneticallymodified crop. For soybeans, the situa-tion is different, since they are virtually100% self-fertilizers and can be planted invery close proximity to other plants withoutfear of horizontal spread. Another optionis the design of transgenic plants that haveonly sterile pollen or – more or less onlyapplicable for greenhouses – completelyprevent cross-pollination by covering theindividual plants. One public fear regards

spreading antibiotic resistance from one(transgenic donor) plant to other wild-type plants or bacteria in the environment.Although prokaryotic promoters for an-tibiotic resistance are sometimes usedin the fabrication and selection of trans-genic constructs, once a transgene hasbeen stably incorporated into the plantgenome, it is under the control of plant(eukaryotic) promoter elements. Hence,antibiotic-resistance genes are unable topass from genetically altered plants intobacteria and remain functional. As statedearlier, another common fear is the cre-ation of a ‘‘super bug.’’ The chance ofcreating a supervirulent virus or bacteriumfrom genetic engineering is unlikely, be-cause the construction of expression cas-settes from viral or bacterial genomesinvolves the removal of the majority ofgenes responsible for the normal functionof these organisms. Even if a resul-tant organism is somewhat functional,it cannot compete for long in naturewith normal, wild-type bacteria of thesame species.

As one can see from the aforementionedsafety strategies, considerable effort isput into the reduction of any potentialrisk from the transgenic plant for theenvironment. In general, the scientificrisk can be kept at a minimum, ifcommon sense is applied – in accordancewith Thomas Huxley (1825–1895) that‘‘Science is simply common sense at itsbest.’’ For example, protein toxins (forvaccine production) should never be grownin food plants.

Additionally, the following can be em-ployed as a kind of risk management toprevent the inappropriate or unsafe use ofgenetically engineered plants:

• An easily recognized phenotypic char-acteristic can be coexpressed in an


engineered product (e.g. tomatoes thatcontain a therapeutic protein can beselected to grow in a colorless varietyof fruit).

• Protein expression can be induced onlyafter harvesting or fruit ripening. Forexample, CropTech’s (Blacksburg, VA,USA) inducible expression system intobacco, MeGA-PharM, leads to veryefficient induction upon leaf injury (har-vest) and needs no chemical inducers.This system possesses a fast inductionresponse and protein synthesis rate, andthus leads to high expression levels withno aged product in the field (no environ-mental damage accumulation).

• Potentially antigenic or immunomodu-latory products can be induced to growin, or not to grow in, a certain plant tis-sue (e.g. root, leaf/stem, seed, or pollen).In this way, for example, farmers can beprotected from harmful airborne pollenor seed dusts.

• Although no absolute system can pre-vent vandalism or theft of the transgenicplants, a very effective, cheap solutionhas been used quietly for many yearsnow in the United States. Plots of thesemodified plants are being grown withabsolutely no indication that they aredifferent from a routine crop. In theMidwest, for example, finding a trans-genic corn plot among the millions ofacres of concurrently growing grain isvirtually impossible. The only questionhere is, if this approach really helpsfacilitating a fair and an open discus-sion with the public. Asking the samequestion for the EU is not relevant:owing to labeling requirements, this ap-proach would not be feasible, as, ingeneral, it is much more difficult toperform open-field studies with trans-genic plants.

5The Way Forward: Moving Plants toHumanlike Glycosylation

As discussed earlier, plant production oftherapeutic proteins has many advantagesover bacterial systems. One very impor-tant feature of plant cells is their capabilityof carrying out posttranslational modifica-tions. Since they are eukaryotes (i.e. have anucleus), plants produce proteins throughan ER (endoplasmatic reticulum) pathway,adding sugar residues also to the pro-tein – a process called glycosylation. Thesecarbohydrates help determine the three-dimensional structures of proteins, whichare inherently linked to their function andtheir efficacy as therapeutics. This glycosy-lation also affects protein bioavailabilityand breakdown of the biopharmaceuti-cal; for example, proteins lacking terminalsialic acid residues on their sugar groupsare often targeted by the immune systemand are rapidly degraded. The glycosyla-tion process begins by targeting the proteinto the ER. During translation of mRNA(messenger RNA) into protein, the ribo-some is attached to the ER, and the nascentprotein fed into the lumen of the ER astranslation proceeds. Here, one set of gly-cosylation enzymes attaches carbohydratesto specific amino acids of the protein.Other glycosylation enzymes either deleteor add more sugars to the core structures.This glycosylation process continues intothe Golgi apparatus, which sorts the newproteins, and distributes them to their finaldestinations in the cell (see Fig. 4). Bacte-ria lack this ability and therefore cannotbe used to synthesize proteins that requireglycosylation for activity. Although plantshave a somewhat different system of pro-tein glycosylation from mammalian cells,the differences usually prove not to be aproblem. Some proteins, however, require


Fig. 4 The glycosylation pathway via ER and Golgi apparatus. In the cytosol,carbohydrates are attached to a lipid precursor, which is then transported into thelumen of the ER to finish core glycosylation. This glycan is now attached to thenascent, folding polypeptide chain (which is synthesized by ribosomes attached tothe cytosolic side of the ER from where it translocates into the lumen) andsubsequently trimmed and processed before it is folded and moved to the Golgiapparatus. Capping of the oligosaccharide branches with sialic acid and fucose is thefinal step on the way to a mature glycoprotein. Source: Dove, A. (2001) Thebittersweet promise of glycobiology, Nat. Biotechnol. 19, 913–917.

Transgenicanimals

Nativeglycoproteins

Bacteria Yeast Transgenicplants

Peptide

Xylose Fucose

Galactose Mannose

N-acetylglucosamine

N-glycolylneuraminic acid

N-acetylneuraminic acid

Fig. 5 Engineering plants to humanlike glycosylation. The first step to achieve humanlikeglycosylation in plants is to eliminate the plant glycosylation pattern, that is, the attachment of β

1–2 linked xylosyl- and α 1–3 linked fucosyl sugars to the protein. Because these two residueshave allergenic potential, the corresponding enzymes Xylosyl- and Fucosyl Transferase areknocked out. In case galactose is relevant for the final product, Galactosyl Transferase is insertedinto the host genome. Galactose is available in the organism so that this single gene insertion issufficient to ensure galactosylation. Source: Kn

..ablein J. (2003) Biotech: A New Era In The New

Millennium – Biopharmaceutic drugs manufactured in novel expression systems,DECHEMA-Jahrestagung der Biotechnologen, Munich, Germany, 21.


humanlike glycosylation (see Fig. 5) – theymust have specific sugar structures at-tached to the correct sites on the moleculeto be maximally effective. Therefore, someefforts are being made in modifying hostplants in such a way that they providethe protein with human glycosylation pat-terns. One example of modifying a plantexpression system in this way is the trans-genic moss, which will be discussed in thenext section.

6Three Promising Examples: Tobacco(Rhizosecretion, Transfection) and Moss(Glycosylation)

To further elaborate on improving glycosy-lation and downstream processing, threeinteresting plant expression systems willbe discussed. All systems share the advan-tage of utilizing nonedible plants (nonfoodand nonfeed) and can be kept in either agreenhouse or a fermenter to avoid anysegregation risk. Another obvious advan-tage is secretion of the protein into themedium so that no grinding or extrac-tion is required. This is very important inlight of downstream processing: proteinpurification is often as expensive as thebiomanufacturing and should never be un-derestimated in the total COGS equation.

6.1Harnessing Tobacco Roots to SecreteProteins

Phytomedics (Dayton, NJ, USA) uses to-bacco plants as an expression system forbiopharmaceuticals. Besides the advantageof being well characterized and used inagriculture for some time, tobacco hasa stable genetic system, provides high-density tissue (high protein production),

needs only simple medium, and can bekept in a greenhouse (see Fig. 6). Opti-mized antibody expression can be rapidlyverified using transient expression assays(short development time) in the plantsbefore creation of transgenic suspensioncells or stable plant lines (longer devel-opment time). Different vector systems,harboring targeting signals for subcellularcompartments, are constructed in paralleland used for transient expression. Apply-ing this screening approach, high express-ing cell lines can rapidly be identified. Forexample, transgenic tobacco plants, trans-formed with an expression cassette con-taining the GFP (Green Fluorescent Pro-tein) gene fused to an aps (amplification-promoting sequence), had greater levels ofcorresponding mRNAs and expressed pro-teins compared to transformants lackingaps. Usually, downstream processing (iso-lation/extraction and purification of thetarget protein) is limiting for such a sys-tem, for example, if the protein has tobe isolated from biochemically complexplant tissues (e.g. leaves), this can be alaborious and expensive process and amajor obstacle to large-scale protein man-ufacturing. To overcome this problem,secretion-based systems utilizing trans-genic plant cells or plant organs asepticallycultivated in vitro would be one solution.However, in vitro systems can be expen-sive, slow growing, unstable, and relativelylow yielding. This is why another inter-esting route was followed. Secretion ofmolecules is a basic function of plantcells and organs in plants, and is espe-cially developed in plant roots. In orderto take up nutrients from the soil, inter-act with other soil organisms, and defendthemselves against numerous pathogens,plant roots have evolved sophisticatedmechanisms based on the secretion ofdifferent biochemicals (including proteins


Root secretion, easy recovery

Greenhouse contained tanks

High density tissue

Salts and water only

Tobacco is well characterized

Stable genetic system

Phytomedics (tobacco):

Fig. 6 Secretion of the biopharmaceuticals viatobacco roots. The tobacco plants are geneticallymodified in such a way that the protein issecreted via the roots into the medium(‘‘rhizosecretion’’). In this example, the tobaccoplant takes up nutrients and water from themedium and releases GFP (Green FluorescentProtein). Examination of root cultivation mediumby its exposure to near ultraviolet-illuminationreveals the bright green-blue fluorescencecharacteristics of GFP in the hydroponic medium

(left flask in panel lower left edge). The picturealso shows a schematic drawing of thehydroponic tank, as well as tobacco plants atdifferent growth stages, for example, callus, fullygrown, and greenhouse plantation. Source:Kn

..ablein J. (2003) Biotech: A New Era in the New

Millennium – Biopharmaceutic DrugsManufactured in Novel Expression Systems,DECHEMA-Jahrestagung der Biotechnologen,Munich, Germany, 21. (See color plate. p. xxv)

like toxins) into their neighborhood (rhi-zosphere). In fact, Borisjuk and coworkerscould demonstrate that root secretion canbe successfully exploited for the continu-ous production of recombinant proteinsin a process termed ‘‘rhizosecretion.’’ Here,an endoplasmic reticulum signal peptideis fused to the recombinant protein, whichis then continuously secreted from theroots into a simple hydroponic medium(based on the natural secretion from rootsof the intact plants). The roots of the to-bacco plant are sitting in a hydroponictank (see Fig. 6), taking up water andnutrients and continuously releasing thebiopharmaceutical. By this elegant set up,downstream processing becomes easy andcost-effective, and also offers the advantage

of continuous protein production that in-tegrates the biosynthetic potential of aplant over its lifetime and might lead tohigher protein yields than single-harvestand extraction methods. Rhizosecretion isdemonstrated in Fig. 6, showing a trans-genic tobacco plant expressing GFP andreleasing it into the medium.

6.2High Protein Yields Utilizing ViralTransfection

ICON Genetics (Halle, Germany) has de-veloped a protein-production system thatrelies on rapid multiplication of viral vec-tors in an infected tobacco plant (seeFig. 7). Viral transfection systems offer


Expression in plant tissue

(d)(c)

(b)

ICON Genetics (tobacco):•

•

•

•

Viral transfectionFast developmentHigh protein yieldsCoexpression of genes

(a)

GFP

CP RbcS

RbcL

Coom assie gel

Fig. 7 Viral transfection of tobacco plants. This new generation platform for fast (1 to 2weeks), high-yield (up to 5 g kg−1 fresh leaf weight) production of biopharmaceuticalsis based on proviral gene amplification in a nonfood host. Antibodies, antigens,interferons, hormones, and enzymes could successfully be expressed with this system.The picture shows development of initial symptoms on a tobacco following theAgrobacterium-mediated infection with viral vector components that contain a GFPgene (a); this development eventually leads to a systemic spread of the virus, literallyconverting the plant into a sack full of protein of interest within two weeks (b). Thesystem allows to coexpress two proteins in the same cell, a feature that allowsexpression of complex proteins such as full-length monoclonal antibodies. Panels(c) and (d) show the same microscope section with the same cells, expressing GreenFluorescent Protein (c) and Red Fluorescent Protein (d) at the same time. The yield andtotal protein concentration achievable are illustrated by a Coomassie gel with proteinsin the system: GFP (protein of interest), CP (coat protein from wild-type virus), RbcSand RbcL (small and large subunit of ribulose-1,5-bisphosphate carboxylase). Source:Kn

..ablein J. (2003) Biotech: A New Era in the New Millennium – Biopharmaceutic Drugs

Manufactured in Novel Expression Systems, DECHEMA-Jahrestagung der Biotechnologen,Munich, Germany, 21. (See color plate. p. xxv)

a number of advantages, such as veryrapid (1 to 2 week) expression time,possibility of generating initial milligramquantities within weeks, high expressionlevels, and so on. However, the existingviral vectors, such as TMV-based vectorsused by, for example, Large Scale BiologyCorp. (Vacaville, CA, USA) for productionof single-chain antibodies for treatmentof non-Hodgkin lymphoma (currently inPhase III clinical trials, see Table 1), hadnumerous shortcomings, such as inabil-ity to express genes larger than 1 kb,inability to coexpress two or more pro-teins (a prerequisite for production of

monoclonal antibodies, because they con-sist of the light and heavy chains, whichare expressed independently and are sub-sequently assembled), low expression levelin systemically infected leaves, and so on.ICON has solved many of these problemsby designing a process that starts withan assembly of one or more viral vec-tors inside a plant after treating the leaveswith agrobacteria, which deliver the nec-essary viral vector components. ICON’sproviral vectors provide advantages of fastand high-yield amplification processes in aplant cell, simple and inexpensive assem-bly of expression cassettes in planta, and


full control of the process. The robustnessof highly standardized protocols allows theuse of inherently the same safe protocolsfor both laboratory-scale as well as indus-trial production processes. In this system,the plant is modified transiently ratherthan genetically and reaches the speed andyield of microbial systems while enjoyingposttranslational capabilities of plant cells.De- and reconstructing of the virus addssome safety features and also increases ef-ficiency. There is no ‘‘physiology conflict,’’because the ‘‘growth phase’’ is separatedfrom the ‘‘production phase,’’ so that nocompetition occurs for nutrients and othercomponents required for growth and alsofor expression of the biopharmaceutical atthe same time.

This transfection-based platform allowsthe production of proteins in a plant host ata cost of US$1 to 10 per gram of crude pro-tein. The platform is essentially free fromlimitations (gene insert size limit, inabilityto express more than one gene) of currentviral vector-based platforms. The expres-sion levels reach 5 g per kilogram of freshleaf tissue (or some 50% of total cellularprotein!) in 5 to 14 days after inoculation.Since the virus process (in addition to su-perhigh production of its own proteins,including the protein of interest) leads tothe shutoff of the other cellular proteinsynthesis, the amount of protein of inter-est in the initial extract is extremely high(Fig. 7). It thus results in reduced costs ofdownstream processing. Milligram quan-tities can be produced within two weeks,gram quantities in 4 to 6 months, andthe production system is inherently scal-able. A number of high-value proteins havebeen successfully expressed, including an-tibodies, antigens, interferons, hormones,and enzymes (see Klimyuk, Marillonnet,Knablein, McCaman, Gleba (2005), books).

6.3Simple Moss Performs ComplexGlycosylation

Greenovation Biotech GmbH (Freiburg,Germany) has established an innovativeproduction system for human proteins.The system produces pharmacologicallyactive proteins in a bioreactor, utilizinga moss (Physcomitrella patens) cell cul-ture system with unique properties (seeFig. 8). It was stated before that posttrans-lational modifications for some proteinsare crucial to gain complete pharmaco-logical activity. Since moss is the onlyknown plant system that shows a highfrequency of homologous recombination,this is a highly attractive tool for produc-tion strain design. By establishing stableintegration of foreign genes (gene knock-out and new transgene insertion) into theplant genome, it can be programmed toproduce proteins with modified glycosyla-tion patterns that are identical to animalcells. The moss is photoautotrophic andtherefore only requires simple media forgrowth, which consist essentially of wa-ter and minerals. This reduces costs andalso accounts for significantly lower in-fectious and contamination risks, but inaddition to this, the system has somemore advantages:

• The transient system allows productionof quantities for a feasibility studywithin weeks – production of a stableexpression strain takes 4 to 6 months.

• On the basis of transient expressiondata, the yield of stable production linesis expected to reach 30 mg L−1 perday. This corresponds to the yield ofa typical fed-batch culture over 20 daysof 600 mg L−1.

• Bacterial fermentation usually requiresaddition of antibiotics (serving as se-lection marker and to avoid loss of the


• Simple medium (photoautotrophic plant needs only water and minerals)

• Robust expression system (good expression levels from 15 to 25°C)

• Secretion into medium via human leader sequence (broad pH range: 4-8)

• Easy purification from low salt medium via ion exchange

• Easy genetic modifications to cell lines

• Stable cell lines / high genetic stability

• Codon usage like human (no changes required)

• Inexpensive bioreactors from the shelf

• Nonfood plant (no segregation risk)

• Good progress on genetic modification of glycosylationpathways (plant to human)

Greenovation (moss system):

Fig. 8 Greenovation use a fully contained moss bioreactor. This company has establishedan innovative production system for human proteins. The system producespharmacologically active proteins in a bioreactor, utilizing a moss (Physcomitrella patens)cell culture system with unique properties. Source: Kn

..ablein J. (2003) Biotech: A New Era in

the New Millennium – Biopharmaceutic drugs Manufactured in Novel Expression Systems,DECHEMA-Jahrestagung der Biotechnologen, Munich, Germany, 21.

expression vector). For moss cultivation,no antibiotics are needed – this avoidsthe risk of traces of antibiotics havinga significant allergenic potential in thefinished product.

• Genetic stability is provided by thefact that the moss is grown in smallplant fragments and not as proto-plasts or tissue cultures avoiding so-maclonal variation.

• As a contained system, the moss biore-actor can be standardized and validatedaccording to GMP standards mandatoryin the pharmaceutical industry.

• Excretion into the simple medium isanother major feature of the moss biore-actor, which greatly facilitates down-stream processing.

As discussed in detail, the first step toget humanlike glycosylation in plants is

to eliminate the plant glycosylation, forexample, the attachment of β-1-2-linkedxylosyl and α-1-3-linked fucosyl sugars tothe protein, because these two residueshave allergenic potential. Greenovationwas able to knockout the relevant glyco-sylation enzymes xylosyl transferase andfucosyl transferase, which was confirmedby RT-PCR (reverse transcriptase PCR).And indeed, xylosyl and fucosyl residueswere completely removed from the glyco-sylation pattern of the expressed proteinas confirmed by MALDI-TOF (matrix as-sisted laser desorption ionization time offlight) mass spectroscopy analysis (seeFig. 9).

A very challenging protein to express isVEGF because this homodimer consistsof two identical monomers linked viaa disulfide bond. To produce VEGF in


Successful knockout of Xylosyl transferase in mossXT-KO plants: RT-PCR and MALDI-TOF analysis

MALDI

Xylosyl transferaseRT-PCR

Control:APS reductase (R10 and R11)

XT114F/XT15R

R10/R11

Fig. 9 Knockout of Xylosyl Transferase in moss. To avoid undesiredglycosylation, greenovation knocked out the Xylosyl and FucosylTransferase, as confirmed by RT-PCR. MALDI-TOF results show thatindeed, xylosyl- and fucosyl-residues were completely removed fromthe glycosylation pattern of the expressed protein (data for knockout ofFucosyl Transferase not shown). Source: Kn

..ablein J. (2003) Biotech: A

New Era in the New Millennium – Biopharmaceutic Drugs Manufacturedin Novel Expression Systems, DECHEMA-Jahrestagung derBiotechnologen, Munich, Germany, 21.

WT

0200400600800

1000

0 50 100 150Fluorescence intensity

Cou

nts

tWTVEGF p31

0200400600800

1000

0 50 100 150Fluorescence intensity

Cou

nts

FACS

analysis

kDa 5 ng 10 ng TPx TPy

SDS

PAGE

Stimulation of human vascular epithelial cells

RPMI -Medium

rh VEGF -Control

(1ng mL−1)

P 27(1ng mL−1)

P 31(2 ng mL−1)

100

110

120

130

Inco

rpor

atio

n ra

te (

% o

f con

trol

)

Biological activity of recombinantVEGF

VEGF121dimer37

26

Fig. 10 Greenovation could successfullyexpress the biopharmaceutical VEGF. Thisgrowth factor is a very complex proteinconsisting of two identical monomers linked viaa disulfide-bond. To produce VEGF in an activeform, the monomers need to be expressed to theright level, correctly folded, assembled, and

linked via the disulfide-bond. The analyticalassays clearly show that expression in mossyielded completely active VEGF. Source:Kn

..ablein J. (2003) Biotech: A New Era in the New

Millennium – from Plant Fermentation to PlantExpression of Biopharmaceuticals, PDAInternational Congress, Prague, Czech Republic.


30 L pilot reactor for moss Two weeks after incubation

Fig. 11 Scaling of photobioreactors up toseveral 1000 L. The moss bioreactor is based onthe cultivation of Physcomitrella patens in afermenter. The moss protonema is grown underphotoautotrophic conditions in a medium thatconsists essentially of water and minerals. Lightand carbon dioxide serve as the only energy andcarbon sources. Cultivation in suspension allows

scaling of the photobioreactors up to several1000 L. Adaptation of existing technology forlarge-scale cultivation of algae is done incooperation with the Technical University ofKarlsruhe. Source: greenovation Biotech GmbH(Freiburg, Germany) and Professor C. Posten,Technical University (Karlsruhe, Germany).

an active form, the following need tobe provided:

• Monomers need to be expressed to theright level.

• Monomers need to be correctly folded.• Homodimer needs to be correctly as-

sembled and linked via a disulfide bond.• Complex protein needs to be secreted in

its active form.

And in fact, all this could be achievedwith the transgenic moss system as shownin Fig. 10. These results are very promisingbecause they demonstrate that this systemis capable of expressing even very complexproteins. In addition to that, the mosssystem adds no plant-specific sugars to theprotein – a major step toward humanlikeglycosylation. Furthermore, moss is arobust expression system leading to highyields at 15 to 25 ◦C and the pH canbe adjusted from 4 to 8 depending on

the optimum for the protein of interest.Adapting existing technology for large-scale cultivation of algae, fermentation ofmoss in suspension culture allows scalingof the photobioreactors up to several1000 L (see Fig. 11). Finally, the medium isinexpensive, since only water and mineralsare sufficient.

7Other Systems Used for Plant Expression

Several different plants have been usedfor the expression of proteins in plants.All these systems have certain advantagesregarding edibility, growth rate, scalability,gene-to-protein time, yield, downstreamprocessing, ease of use, and so on, whichI will not discuss in further detail here. Aselection of different expression systemsis listed:


Alfalfa Ethiopianmustard

Potatoes

Arabidopsis Lemna RiceBanana Maize SoybeanCauliflower Moss TomatoesCorn Oilseeds Wheat

Some of these systems have been usedfor research on the basis of their easeof transformation, well-known characteri-zation, and ease to work with. However,they are not necessarily appropriate forcommercial production. Which crop is ul-timately used for full-scale commercialproduction will depend on a number offactors including

• time to develop an appropriate system(gene-to-protein);

• section of the plant expressing theproduct/possible secretion;

• cost and potential waste productsfrom extraction;

• ‘‘aged’’ product/ease of storage;• long-term stability of the storage tissue;• quantities of protein needed (scale

of production).

Depending on the genetic complexityand ease of manipulation, the develop-ment time to produce an appropriatetransgenic plant for milligram productionof the desired protein can vary from 10to 12 months in corn as compared to onlyweeks in moss. Estimates for full GMPproduction in corn are 30 to 36 monthsand approximately 12 months for moss.Expression of the protein in various tissuesof the plant can result in a great variation inyield. Expression in the seed can often leadto higher yields than in the leafy portion ofthe plant. This is another explanation forthe high interest in using corn, which hasa relatively high seed-to-leaf ratio. Extrac-tion from leaf can be costly as it contains a

high percentage of water, which could re-sult in unavoidable proteolysis during theprocess. Proteins stored in seeds can bedesiccated and remain intact for long peri-ods of time. The purification and extractionof the protein is likely to be done by adapta-tions of current processes for the extractionand/or fractionation. For these reasons, itis anticipated that large-scale commercialproduction of recombinant proteins willinvolve grain and oilseed crops such asmaize, rice, wheat, and soybeans. On thebasis of permits for open-air test plotsissued by the USDA for pharmaceuticalproteins and industrial biochemicals, cornis the crop of choice for production with73% of the permits issued. The other majorcrops are soybeans (12%), tobacco (10%),and rice (5%).

In general, the use of smaller plants thatcan be grown in greenhouses is an effectiveway of producing the biopharmaceuticalsand alleviating concerns from environ-mental activist groups that the transgenicplant might be harmful to the environ-ment (food chain, segregation risk, geneticdrift, etc.).

8Analytical Characterization

Validated bioanalytical assays are essentialand have to be developed to characterizethe biopharmaceuticals during the produc-tion process (e.g. in-process control) andto release the final product for use as adrug in humans. These assays are appliedto determine characteristics such as pu-rity/impurities, identity, quantity, stability,specificity, and potency of the recombi-nant protein during drug development.Since the very diverse functions of dif-ferent proteins heavily depend on theirstructure, one very valuable parameter in


protein characterization is the elucidationof their three-dimensional structure. Al-though over the last couple of years a lot ofeffort was put into a method for improv-ing the elucidation of protein structures(during my PhD thesis, I was also work-ing in this fascinating field together withmy boss Professor Robert Huber, NobelPrize Laureate in 1988, ‘‘for the determi-nation of the three-dimensional structureof a photosynthetic reaction centre’’), it isstill very time consuming to solve the 3-Dstructure of larger proteins. This is why de-spite the high degree of information thatcan be obtained from the protein struc-ture, this approach cannot be applied ona routine basis. Therefore, tremendous ef-forts are put into the development of otherassays to guarantee that a potent biophar-maceutical drug is indeed ready for usein humans.

9Conclusion and Outlook

The production of protein therapeuticsfrom transgenic plants is becoming areality. The numerous benefits offeredby plants (low cost of cultivation, highbiomass production, relatively fast gene-to-protein time, low capital and operatingcosts, excellent scalability, eukaryotic post-translational modifications, low risk ofhuman pathogens, lack of endotoxins,as well as high protein yields) virtu-ally guarantee that plant-derived proteinswill become more and more commonfor therapeutic uses. Taking advantage ofplant expression systems, the availabilityof cheap protein-based vaccines in un-derdeveloped countries of the world ispossible in the near future. The cost ofvery expensive hormone therapies (ery-thropoietin, human growth hormone, etc.)

could fall dramatically within the nextdecade because of the use of, for exam-ple, plant expression systems. Fears aboutthe risks of the plant expression technol-ogy are real and well founded, but with adetailed understanding of the technology,it is possible to proactively address thesesafety issues and create a plant expres-sion industry almost free of mishaps. Forthis purpose, the entire set up, consistingof the specific plant expression system andthe protein being produced, needs to be an-alyzed and its potential risks assessed on acase-by-case basis. As plant-derived thera-peutics begin to demonstrate widespread,tangible benefits to the population and asthe plant expression industry develops alonger safety track record, public accep-tance of the technology is likely to improvecontinuously. Plants are by far the mostabundant and cost-effective renewable re-source uniquely adapted to complex bio-chemical synthesis. The increasing cost ofenergy and chemical raw materials, com-bined with the environmental concernsassociated with conventional pharmaceu-tical manufacturing, will make plants evenmore compatible in the future. With thewords of Max Planck (1858–1947) ‘‘Howfar advanced Man’s scientific knowledgemay be, when confronted with Nature’simmeasurable richness and capacity forconstant renewal, he will be like a mar-veling child and must always be preparedfor new surprises,’’ we will definitely dis-cover more fascinating features of plantexpression systems. But there is no needto wait: combining the advantages of sometechnologies that we already have in handcould lead to the ultimate plant expressionsystem. This is what we should focus on,because, then, at the dawn of this new mil-lennium, this would for the first time yieldlarge-enough amounts of biopharmaceuti-cals to treat everybody on our planet!


Acknowledgments

I would like to thank the companies green-ovation Biotech GmbH (Freiburg, Ger-many), ICON Genetics (Halle, Germany),and Phytomedics (Dayton, NJ, USA) forproviding some data and figures to preparethis manuscript.

See also Bioprocess Engineering;Expression Systems for DNA Pro-cesses; Plant Gene Expression,Regulation of.

Bibliography

Books and Reviews

Fischer, R., Schillberg S. (Eds.) (2004) MolecularFarming: Plant-made Pharmaceuticals andTechnical Proteins, Wiley, ISBN: 3-527-30786-9

Fischer, R., et al. (2004) Plant-based productionof biopharmaceuticals, Curr. Opin. Plant Biol.7(2), 152–158.

Horn, M.E., Woodard, S.L., Howard, J.A. (2004)Plant molecular farming: systems andproducts, Plant Cell Rep. 22(10), 711–720.

Klimyuk, V., Marillonnet, S., Knablein, J., Mc-Caman, M., Gleba, Y. (2005) Production ofRecombinant Proteins in Plants, in: ModernBiopharmaceuticals – Design, Development andOptimization, Wiley-VCH, in press.

Knablein, J. (2004) Biopharmaceuticals Ex-pressed in Plants – A New Era in the NewMillennium, in: Muller, R., Kayser, O. (Eds.)Applications in Pharmaceutical Biotechnology,Wiley-VCH, ISBN 3-527-30554-8.

Ma, J.K., Drake, P.M., Christou, P. (2003) Theproduction of recombinant pharmaceuticalproteins in plants, Nat. Rev. Genet. 4(10),794–805.

Stoger, E., et al. (2004) Antibody production intransgenic plants, Methods Mol. Biol. 248,301–318.

Primary References

Arakawa, T., Chong, D.K.X., Langridge, W.H.R.(1998) Efficacy of a food plant-based oral

cholera toxin B subunit vaccine, Nat.Biotechnol. 16, 292–297.

Arakawa, T., Yu, J., Chong, D.K., Hough, J.,Engen, P.C., Langridge, W.H. (1998) A plant-based cholera toxin B subunit-insulin fusionprotein protects against the development ofautoimmune diabetes, Nat. Biotechnol. 16,934–938.

Arthur D. Little, Inc. (ADL), AgIndustriesResearch, Cambridge, MA, Copyright 2002.

Artsaenko, O., et al. (1998) Potato tubers as abiofactory for recombinant antibodies, Mol.Breeding 4, 313–319.

Beachy, R.N., Fitchen, J.H., Hein, M.B. (1996)Use of Plant Viruses for Delivery of VaccineEpitopes, in: Collins, G.B., Sheperd, R.J. (Eds.)Engineering Plants for Commercial Products andApplications, New York Academy of Sciences,New York, pp. 43–49.

Boothe, J.G., Parmenter, D.L., Saponja, J.A.(1997) Molecular farming in plants: oilseeds asvehicles for the production of pharmaceuticalproteins, Drug Develop. Res. 42, 172–181.

Borisjuk, N.V., Raskin, I., et al. (1999) Produc-tion of recombinant proteins in plant rootexudates, Nat. Biotechnol. 17, 466–469.

Borisjuk, N.V., Raskin, I., et al. (2000) Tobaccoribosomal DNA spacer element stimulatesamplification and expression of heterologousgenes, Nat. Biotechnol. 18, 1303–1306.

Cabanes-Macheteau, M., et al. (1999) N-glycosylation of a mouse IgG expressedin transgenic tobacco plants, Glycobiology 9,365–372.

Chance, R.E., Frank, B.H. (1993) Research,development, production and safety ofbiosynthetic human insulin, Diabetes care16(3), 133–142.

Chaudhary, S., Parmenter, D.L., Moloney, M.M.(1998) Transgenic Brassica carinata as avehicle for the production of recombinantproteins in seeds, Plant Cell Rep. 17, 195–200.

Conrad, U., Fiedler, U. (1994) Expression ofengineered antibodies in plant cells, Plant Mol.Biol. 26, 1023–1030.

Conrad, U., Fiedler, U., Artsaenko, O., Phillips,J. (1998) High-level and stable accumulationof single-chain Fv antibodies in plant storageorgans, J. Plant Physiol. 152, 708–711.

Cramer, C.L., Boothe, J.G., Oishi, K.K. (1999)Transgenic plants for therapeutic proteins:linking upstream and downstream strategies,Curr. Top. Microbiol. Immunol. 240, 95–118.


Cramer, C.L., et al.. (1996) Bioproduction ofHuman Enzymes in Transgenic Tobacco, in:Collins, G.B., Sheperd, R.J. (Eds.) EngineeringPlants for Commercial Products andApplications, New York Academy Of Sciences,New York, 62–71.

Dalsgaard, K., et al. (1997) Plant-derived vaccineprotects target animals against a viral disease,Nat. Biotechnol. 15, 248–252.

Davies, L., Plieth, J. (2001) The challenge ofmeeting the escalating demand for proteins,Scr Mag 10, 25–29.

Della-Cioppa, G., Grill, L.K. (1996) Productionof Novel Compounds in Higher Plants byTransfection with RNA Viral Vectors, in:Collins, G.B., Sheperd, R.J. (Eds.) EngineeringPlants for Commercial Products andApplications, New York Academy of Sciences,New York, pp. 57–61.

Dieryck, W., et al. (1997) Human haemoglobinfrom transgenic tobacco, Nature 386, 29–30.

Doran, P.M. (2000) Foreign protein productionin plant tissue cultures, Curr. Opin. Biotechnol.11, 199–204.

Dove, A. (2001) The bittersweet promise ofglycobiology, Nat. Biotechnol. 19, 913–917.

Dove, A. (2002) Unkorking the biomanufactur-ing bottleneck, Nat. Biotechnol. 20, 777–779.

Drake, P.M., Chargelegue, D., Vine, N.D., VanDolleweerd, C.J., Obregon, P., Ma, J.K. (2002)Transgenic plants expressing antibodies: amodel for phytoremediation, FASEB J. 16(14),1855–1860.

Drug & Market Development Publications,Antibody Engineering: Technologies, Applicationsand Business opportunities, Westborough, MA,Copyright 2003.

Evangelista, R.L., Kusnadi, A.R., Howard, J.A.,Nikolov, Z.L. (1988) Process and economicevaluation of the extraction and purification ofrecombinant glucouronidase from transgeniccorn, Biotechnol. Prog. 14, 607–614.

Fischer, R., Emans, N. (2000) Molecular farmingof pharmaceutical proteins, Transgenic Res. 9,279–299.

Fischer, R., Hoffmann, K., Schillberg, S.,Emans, N. (2000) Antibody production bymolecular farming in plants, J. Biol. Regul.Homeost. Agents 14, 83–92.

Ganz, P.R., et al.. (1996) Expression of HumanBlood Proteins in Transgenic Plants: TheCytokine GM-CSF as a Model Protein,in: Owen, M.R.L., Pen, J. (Eds.) TransgenicPlants: A Production System for Industrial and

Pharmaceutical Proteins, John Wiley & Sons,London, UK, 281–297..

Garber, K. (2001) Biotech industry faces newbottleneck, Nat. Biotechnol. 19, 184–185.

Giddings, G., Allison, G., Brooks, D., Carter, C.(2000) Transgenic plants as factories forbiopharmaceuticals, Nat. Biotechnol. 18,1151–1155.

Goddijn, O.J.M., Pen, J. (1995) Plants asbioreactors, Trends Biotechnol. 13, 379–387.

Hamamoto, H., et al. (1993) A new tobaccomosaic virus vector and its use forthe systematic production of angiotensin-I-converting enzyme inhibitor in transgenictobacco and tomato, Biotechnology 11,930–932.

Hiatt, A., Cafferkey, R., Bowdish, K. (1989)Production of antibodies on transgenic plants,Nature 342, 76–78.

Hood, E.E., Jilka, J.M. (1999) Plant-basedproduction of xenogenic proteins, Curr. Opin.Biotechnol. 10, 382–386.

Johnson, E. (1996) Edible plant vaccines, Nat.Biotechnol. 14, 1532–1533.

Knablein, J. (2003) Biotech: A new era inthe new millennium – fermentation andexpression of biopharmaceuticals in plants,SCREENING – Trends Drug Discov 4, 14–16.

Knablein, J., McCaman, M. (2003) Modernbiopharmaceuticals-recombinant protein ex-pression in transgenic plants, SCREEN-ING – Trends Drug Discov 6, 33–35.

Knablein, J., Huber, R., et al. (1997)[Ta6Br12]2+, a tool for phase determinationof large biological assemblies by X-ray crystal-lography, J. Mol. Biol. 270, 1–7.

Kumagai, M.H., et al. (1993) Rapid, high-level expression of biologically active alpha-trichosanthin in transfected plants by an RNAviral vector, Proc. Natl. Acad. Sci. USA 90,427–430.

Kusnadi, A., Nikolov, Z.L., Howard, J.A. (1997)Production of recombinant proteins intransgenic plants: practical considerations,Biotechnol. Bioeng. 56, 473–484.

Ma, J.K.C. (2000) Genes, greens, and vaccines,Nat. Biotechnol. 18, 1141–1142.

Ma, J.K.C., Hein, M.B. (1995) Plant antibodiesfor immunotherapy, Plant Physiol. 109,341–346.

Ma, J.K.C., Hein, M.B. (1996) AntibodyProduction and Engineering in Plants, in:Collins, G.B., Sheperd, R.J. (Eds.) EngineeringPlants for Commercial Products and


Applications, New York Academy of Sciences,New York, pp. 72–81.

Ma, J.K.C., Hiatt, A. (1996) Expressing Anti-bodies in Plants for Immunotherapy, in:Owen, M.R.L., Pen, P. (Eds.) Transgenic Plants:A Production System for Industrial and Pharma-ceutical Proteins, John Wiley & Sons, London,UK, pp. 229–243.

Ma, J.K.C., Vine, N.D. (1999) Plant expressionsystems for the production of vaccines, Curr.Top. Microbiol. Immunol. 236, 275–292.

Ma, J.K.C., et al. (1998) Characterization ofa recombinant plant monoclonal secretoryantibody and preventive immunotherapy inhumans, Nat. Med. 4(5), 601–606.

Ma, S.W., et al. (1997) Transgenic plantsexpressing autoantigens fed to induce oralimmune tolerance, Nat. Med. 3, 793–517.

McCormick, A.A., et al. (1999) Rapid productionof specific vaccines for lymphoma byexpression of the tumor-derived single-chainFv epitopes in tobacco plants, Proc. Natl. Acad.Sci. USA 96, 703–708.

McGarvey, P.B., et al. (1995) Expression ofthe rabies virus glycoprotein in transgenictomatoes, Biotechnology 13, 1484–1487.

Moloney, M.M. (1995) ‘‘Molecular farming’’in plants: achievements and prospects,Biotechnol. Eng. 9, 3–9.

Morrow, K.J. (2002) Economics of antibodyproduction, Genet. Eng. News 22, 34–39.

Mushegian, A.R., Shepard, R.J. (1995) Geneticelements of plant viruses as tools for geneticengineering, Microbiol. Rev. 59, 548–578.

Parmenter, D.L., et al. (1995) Production ofbiologically active hirudin in plant seedssing oleosin partitioning, Plant Mol. Biol. 29,1167–1180.

Pen, J. (1996) Comparison of Host Systemsfor the Production of Recombinant Proteins,in: Owen, M.R.L., Pen, J. (Eds.) TransgenicPlants: A Production System for Industrial andPharmaceutical Proteins, John Wiley & Sons,London, UK, pp. 149–167.

Ponstein, A.S., Verwoerd, T.C., Pen, J. (1996)Production of Enzymes for IndustrialUse, in: Collins, G.B., Sheperd, R.J. (Eds.)Engineering Plants for Commercial Products andApplications, Vol. 792, New York Academy ofSciences, New York, pp. 91–98.

Raskin, I., Fridlender, B., et al. (2002) Plants andhuman health in the twenty-first century,Trends Biotechnol. 20, 522–531.

Richter, L.J., Thanavala, Y., Arntzen, C.J., Ma-son, H.S. (2000) Production of hepatitisB surface antigen in transgenic plantsfor oral immunization, Nat. Biotechnol. 18,1167–1171.

Ruggiero, F., et al. (2000) Triple helix assemblyand processing of human collagen producedin transgenic tobacco plants, FEBS Lett. 469,132–136.

Sijmons, P.C., et al. (1990) Production ofcorrectly processed human serum albumin intransgenic plants, Biotechnology 8, 217–221.

Smith, M.D. (1996) Antibody production inplants, Biotechnol. Adv. 14, 267–281.

Smith, M.D., Glick, B.R. (2000) The productionof antibodies in plants, Biotechnol. Adv. 18,85–89.

Stoger, E., et al. (2000) Cereal crops asviable production and storage systems forpharmaceutical scFv antibodies, Plant Mol.Biol. 42, 583–590.

Tacket, C.O., Mason, H.S. (1999) A review oforal vaccination with transgenic vegetables,Microbes Infect. 1, 777–783.

Tacket, C.O., et al. (1998) Immunogenicity inhumans of a recombinant bacterial antigendelivered in a transgenic potato, Nat. Med. 4,607–609.

Technology Catalysts International Corporation,Biopharmaceutical Farming, Falls Church, VA,Copyright 2002.

Thanavala, Y., et al. (1995) Immunogenicity oftransgenic plant-derived hepatitis B surfaceantigen, Proc. Natl. Acad. Sci. USA 92,3358–3361.

The Context Network, BiopharmaceuticalProduction in Plants, Biopharma Prospectus,West Des Moines, IA, Copyright 2002.

Tomsett, B., Tregova, A., Garoosi, A., Cad-dick, M. (2004) Ethanol-inducible gene expres-sion: first step toward a new green revolution?Trends Plant Sci. 9(4), 159–161.

Valdes, R., et al. (2003) Hepatitis B surfaceantigen immunopurification using a plant-derived specific antibody produced in largescale, Biochem. Biophy. Res. Commun. 310,742–747.

Vandekerckhove, J., et al. (1989) Enkephalinesproduced in transgenic plants using modified2S storage proteins, Biotechnology 7, 929–932.

Whitelam, G.C. (1995) The production ofrecombinant proteins in plants, J. Sci. FoodAgric. 68, 1–9.


Whitelam, G.C., Cockburn, W. (1996) Antibodyexpression in transgenic plants, Trends PlantSci. 1, 268–272.

Zhong, G.Y., et al. (1999) Commercial produc-tion of aprotinin in transgenic maize seeds,Mol. Breeding 5, 345–356.

Documents

Plant-based Expression of Biopharmaceuticals