Journa l of Scientific & Industrial Research

Vol. 60, October 200 I, pp 773-778

Large Scale Processing of Tetanus Toxin from Fermentation Broth S D Ra vetkar ' , S 8 Rah alkar and C G Kulkarni

Serum Institute of India Research Found:1tion, 2 12/2, Hadapsar, Punc 4 11 02~. India

Ph: lJ 1-20-6993900: fax: lJ 1-20-6993921

Received: I January 200 I : accepted: 2 1 June 200 I

Separation of bacterial cell s, conccntrat ion and stcri lc fi I ration arc some of the m:1jor unit operations in the processing of large volumes of ferment at ion broth in the producti on of bacterin ! vaccines. Separation of cells is traditionall y :Jchicvcd by ccntri fugation or by dead-ended depth filtrat ion. These so lid-liquid scpMationmcthods arc time consuming, tox in recoveries ;lrc not sat isfactory and it is diffi cult to validate a depth filtration system. Tctnnus toxoid i s prepared by detoxifying the cu lture filtrates of Clostridiu111 teta11i :1nd further purifying it by ultrafi ltrntion and s;1 lt fr;1ctinnntion. This nrtic lc desc ribes compari son of performance of dead­end filt rat ion and Tangentia l Flow Filtration (TFf<) system for cln rifi c;l ti on of tct;mus fcrmcnt :Jt ion broth and further usc of ;1 simi l:~r system forconccntr:J tion ;111d purification oftctnnus toxoid .

Introduct ion

Immunizat ion is a signifi cant component of Primary Hea lth Care and late ly lot of emph as is has been placed on immuni zat ion as a hea lth interve nti on and has ultimately tak en the shape of Uni versal lmmuniza ti on Program. DPT (diphth eria , pertu ss is and tetanus) is co mbinati on vacc in es, whi ch hold brig ht fu ture for immunizat ion program . DPT group of vacc ines form a major part or vacci nes again st th e " targe t" d iseases identified by the immunizati on program. Presentl y, there are several manufacturers of OPT group or vacc ines. lt should be rea li zed tha t th e tec hn ology o r vaccine production is changing at a fa st pace and . therefore, the ex isting units should generate enough surp lu s to in vest in R&D- not onl y fo r products a lone , but fo r the manufacturing methods also.

One area of development s is to produce a purer produc t. l t will red uce th e incidences or unt owa rd reacti ons, after immuni zation , as the react ions are often caused by the presence of non-speci fie material s. i e. impurities. Purer anti gens would also result in better qua lity co mbinati on vacc ines with less sup rressive interacti on with other antigens.

Separation of bacterial ce ll s, concentrat ion and :-; tcri le filtrat ion of tox ins are some of the major uni t operati ons

· CorrcsronJi ng author

in the processing of large volumes of fe rmentat ion media in the prod ucti on of bacteri al vacc ines. The separati on or ce ll s is the most importan t step, which is achi eved traditi onall y by ccntri fu gati on or by the dead-ended depth filtra tion method. Such so li d- li quid separati on methods are time con sumin g, and toxin recove ri es a re not sati sfactory. Furthermore, it is difficult to va lidate a depth filtrati on system. whi ch is mandatory under GMP (good manufac turing pract ices).

Me mbrane filters sepa rat e com pone nts and suspensions on the bas is of s ize. In tongcntiol.flmr

.filtmtion (TFF) mode, particles or so lutes retain ed by membrane arc con tinuously removed in the retcn tate fl ow ing tangentia ll y across the membrane surface. The clarified so luti on flows through the membrane into the permeate, also ca ll ed the cmssflowflltmtion .

Tetanu s toxo id is prepared by detoxify in g the cultu re filt rates of Clostridiu111 tetoni, and furth er purifying it by ultra filtrat ion and sa lt fractionatio n. Sati sfactory remova l of ce ll debris (i e, clar ifi ca ti on) is critica l to obtain good yiclds.lnthe conventi onal met hod. the dead­end filtrati on, incorporating filter pads or sui table pore sizes, the separati on must provide co mplete retention or ce ll s, and maximal passage and recovery of so lu ble end­product . When tetanus fermentati on broth is clarified. the large bacterial mass blocks most of the dead ended filtration pads req uiring still hi gher inl et pressures or change of pads altogether. The cla ri fi cd tox in so lution is

774 J SCIIND RES VOL 60 OCTOBER 2001

then polished by pass in g it through pad filters to produce sterile tetanus tox in . Also, the depth filtration is not a closed process (due to the need to change pads during the process) and the working personnel are exposed to the toxin. Additionally, the pads also absorb the active material, resulting in reduced yields.

Another GMP aspect observed is the inability to tes t integrity of pads and poss ible leaching of extractable matter. For these reasons, an alternative technology was looked at seriously, that could be validated, and the trial s were conducted.

A microporous TFF system provides a practica l and economica l alternative to the dead ended filtrati on1

• It is a closed system which reduces the exposure of personnel to the tetanus culture. The entire operation conforms to th e GMP norms and can be validated as and when required2. Validation is a key step in th e ove rall manufac turing process. It is defined as the procedure of obtaining and documenting evidence of the performance of a system, which adequately demonstrates compliance with the design criteria.

The study compares the conventional depth filtration method with that of microporous tangential fl ow filtration (TFF), es tabli shes parameters that opt imi ze overa ll concentrat ion, di afiltration and result in improvement in the preparation of tetanus vaccine. It also describes aspects of validation of the TFF system.

Materials and Methods

The Harvard strain of Clostridium tetani (from the Central Research Institute, Kasauli , India, the national control authority) was used for the producti on of tetanus toxin in fermenter vessels ( 1000 L). Muell er and Miller (M&M) medium' was used for growing C/. tetani . The fermentation vessel had a working capacity of I 000 L. Sterilization of the med ium was carried out by steam at 12 I oc for 30 min and after cooling it down to 35 oc, it was inoculated with a I d old seed and incubated for 6 d at 35 oc under continuous mild agitat ion.

At the end of the incubation period, samples were drawn for purity checks and estimati on of anti ge n con centration. Th e purit y was c hecked by : ( i) microscopic observation, wherein on ly the Gram-posit ive bacill i should be visible, ( ii ) by inoculating nutrient agar broth/slants and incubating these fo r 2 d, wherein no growth shou ld be visible.

The ant igen concentrati on was determin ed by th e class ical flocculati on meth od (Ram on Fl occ uiJti on)

wherein the sa mple is fl occulated against a standard antitoxi n of known strength (Lf/mL).

If the fermented broth is fou nd to be pure and of sati sfactory concentrat ion of the tox in (Lf/mL ), it is subjected to further filtration and processin g as detailed below.

A: Dead-Ended Depth Filtration

Filter Material

Filter pads, 20 em x 20 em, 3-5 11 pore size (K-200 pads, Seitz Filter Werke Bad Kreuznach , West Germany);

Filter Press Assemblies (S traussburger, West Germany)

The press assemb ly holds 24 pads.

Operating Conditions

The operations were carried out at 34-35 oc under I 2 ps1 pressure.

Performance

The output was 80-100 L filtrate in 50-60 min and required 300-400 filter pads to filter 1000 L broth. The output thus obtained was subjected to another depth filtration using EKS pads, 20 x 20 em, and 24 pads were required to obta in a sterile product.

The procedure detai led required change of pads during the process. A sli ght variation in inlet pressure results in a turbid output , needing yet another clari fication cyc le. This problem was encountered in 18 out of 52 batches processed in a year.

B: Tangential Flow Filtration

It is essential to select a su itab le TFF dev ice taking into account the criti ca l aspects o f ce ll recove ry, possibility of shearing of cell s within the dev ice, control of temperature, repeated usage, possibility of steami ng, etc. fn order to find a suitable solution to above problems, a set of tria ls was conducted on a pil ot unit of TFF system (Mi ll ipore Ltd ) wi th hydrophi lic PVDF membrane of 0.22 ~~po re si ze. whi ch has a very low prote in bi nding characteristi cs. It constituted the vali datab le devi ce for biopha rmaceut ical applicati ons .

Fi ve batches were compared with cla rification sluuies on Prostak Pilot unit. Subsequent ly, production scale ba tch es were tak en up and th e sy st e m wa s we ll standardi zed for I 000 L batches.

RAVETKAR 1'1 a/.: PROCESS I G OF TETANUS TOX IN 775

System Description

'M illipore Prostak Demo Unit ' was selected for the clarificati on of tetanus toxoid broth . The unit has a screw pump con nected to a variab le drive. Flow rate of the pump can be varied from 0 to 30 Llmin. A pressure gauge was in sta lled on the feed line and another one on the retentate line. A sanitary type diaphragm valve was also installed on the retentate line, which is used in controlling the trans-membrane pressure, ac ross the membrane.

Prostak modules (Catalogue o PSDVAG021. 0.22f.lm, 2 ft2 ) were used for the trials. Five modules were installed in a series. The feed entered th e first module through the feed manifo ld of the Prostak holder and retentate of the first module entered as a feed to the second module . The arrangement conti nued and the final retentate came out from the fifth module. The permeate from first module entered into the permeate channel of the second modul e, and from second to third it flows up to the fifth modul e. It is co ll ected after the fift h module .

The trials were conducted to remove cell debris from fermenta tion broth and sim ultaneously clari fying the tetanus toxin using Durapore Prosta h: Modu le of O.n p pore size. The modules we re se lected for the ce ll clarification step due to their open channel design as there would be hi gher chances of membrane fouling in case of thin channe l design. For the pi lot studi es, I 00 L cultures, each from five successive batches , were used for filtration on pilot-scale for TFF system. The toxin was processed through the Prostak mod ules. Diafiltration was clone with normal sa line. Five 20ft" mod ules were used for trials and the system was operated at a cross flow of 2000 L/h at an operating temperature of n C. The operat ing parameters were optimi zed based on rl ow and pressure excursion experiments .

After the filtration of the toxin was comp lete, the furt her processing involved detoxifi cation , and 0.45 per cent (v/v) fo rmalin was added for the purpose, and after adjusting pH to 7.6±0.5 , it was incubated at 35±1 oc for 28 d. After completion of detoxification , the toxoid was checked for its spec ific toxicity. Once the toxoid passed the specific toxicity test, it was concentrated using the

Pell icon system, described subsequently.

Validation of the TFF System

The selection of a membrane, as specified earlier. was mainly guided by the possibility of validat ion of the

sys te m. Validation of th e system includ ed: ( i) determination of the pre- and post-operation integrity testing of the membrane dev ice, which ensured repeated use, (ii ) removal of the storage agent, i e, fo rmalin from the modules. Formalin was removed by flushin g the system with distilled water. Samples were co llected at different intervals. To 3 mL sample, 0.5 mL chromotropic acid was added and the mi xture was placed in boiling water bath fo r 5 min. The presence of formalin was confirmed by th e deve lopme nt of pink co lour, ( iii ) check ing absence of C/. tetoni- as the membrane used was of pore size 0.22 p , the absence of C/. tetoni from the c larifi ed tox in needed to be va lidated. Th is was carri ed out by passing a representative sample of 250 mL through the GVWP disc (Durapore PVDF 0.22~t disc) , which was placed on the surface of M&M agar. On another M & M agar plate, overnight grown culture of C/. tetani in fluid thioglycollate medi um was streaked as a positive controL (iv) sanitization of the system­for sani ti zing the system, chemica ls such as acetic acid and sod ium hypochlorite \vcre used . The cleaning procedure ensured removal of th ese c lean ing agents. Removal of aceti c acid (below I 0 ppm) was confirmed by comparing seria ll y d iluted sample of acetic acid with methyl reel indi cator. In case of sodium hypochlorite. procedure similar to that adopted in case of acetic ac id was used, a universal indicator was used instead of methyl red indicator.

Concentration and Ammonium Sulphate Fractionation of Toxoid Using Pellicon System

After detox ifi cat ion and specific toxici ty testing. the next important aspect was the selection of ultrafilration membrane for concentrating the toxoid. The selection of the device was based on the criteria of void-free character istics. integ rity testab le, lowe r binding to pro tein s and hi g he r flux. Jt also offe red g reater mechanical strength and offerrecl fl exibility in cleaning. The selection of Biomax from Millipore of modified Polycthylsu lphon e (PES) satisfied the cr it ical requirement of the cassette se lection . The se lection of thin channel configuration enhanced the flux aspect in addition to the benefit s offered by the basic Biomax membrane. There was no problem of fouling due to cc li s as it was primarily an application in volving dissolved molecules of proteins. The batch size at this stage of concentration of tetanus tox oid was in the range of I 040-1060 L. The molecular wei ght of tetanus toxoid is arou nd

776 J SCI IND RES VOL 60 OCTOBER 2001

I 50,000 daltons, which inc ludes the monomer, dime rs and oligomers of vari ous sizes. The area of the device Biomax I OK NMWL cassettes was 50 ft2

. The cassettes were tested for integrity, before and after every batch and the values were recorded .

The diafiltration process was performed afte r the second fracti on of ammonium sulph ate prec ipitati on of toxoid, a purification step for sulphate removal. Thi s was carried out with normal sa line and in volved I 0-12 wash

cycles.

Results and Discussion

The summary of the pilot-scale trial s is gtven tn Table I . The trial s showed bette r recovery and a lso effective and economic clarification of the te tanus toxin fermentation due to good flux . The data were scaled up for process ing a I 000 L batch and also processed with the commercial-scale data on the othe r unit operation. The detail s of the system in pilot- and plant-scales are given in Table 2.

Table !-Summary of Pilot scale tri als

Parameter Batch No.

2 3 4 5 Vi, L 100 100 100 100 100 Yr. L 5.1 2.0 2.0 3.0 2.0 Yp, L 102 102 105 106 106 Yd , L 10 8 10 10 10 Time, min 180 160 170 IRO ISO Flux, Lmh 58 56 40 3R 3R ATP, bar 1.2 10. 1.3 1.3 1.3 Initial, U/mL 75 65 90 110 100 Permeate, LflmL 70 60 80 100 95 Protein recovery,

per cent 95.2 94. 1 93 .3 96.0 95 .0

Vi - Init ia l volume. Yr - Rctcntate volume, Yp - Permeate volu me, Yd- Diafilration volume ATP- Average trans-membrane pressure Flux -the permeate flow is described in un its of liters/uni t area/ unit time Lt/mL- Unit to describe the anti gen concentrat ion

Table 2-Details of system in Pilot and Plant Scale

Parameter Pilot sca le Plant scale

Batch vo lume, L 100 1000

Filtration area per module, ft ~ 2 20

Total Filtration area, 1'! 2 10 100 Transmembrane pressure 1.0-1.5 1.0 - 1.5 of operat ion. bar

Protein recovery, per cent 95 - I 00 95- 9R

Table 3-Parameters monitored during orc ration of rbnt

sca le system

Parameter Batch No.

1467 1469 1479 1481 1488

Yi , L 1000 1000 1000 1000 1000

Yr. L 26 25 25 20 30

Yr. L 1040 1020 1035 1040 1050

Yd , L 60 60 60 60 80

Time, min 300 300 450 305 315

Flux, Lrnh 22.4 2 1.9 14.8 22.0 21.5

ATP, bar 1.2 1. 2 1.3 1.3 u Initi al, Lt/mL 80 90 110 II 0 100

Permeate, LflmL 90 80 100 11 0 90 Protei n recovery.

per cent 92.4 90.6 94.0 94.5 95.0

Vi - Initi al volume, Yr - Retentate volume, Yr - Permeate volume. Yd- Diafilration volume ATP - Average trans-membrane pressure Flux- the permeate tlow is described in uni ts of liters/unit area/

unit time Lf/mL- Unit to describe the antigen concentrati on

Table 4--Comrarison of dead-end filtrat ion with tangential

tlow fi lt ration Parameter Dead-end Tangentia l

filtra tion tlow filtrat ion Filtration time, h 6-7 5 Sanitization/Cleaning ti me, h 6 1.5 h

Total working time, h 12 h, minimum 6.5 h

Product recovery, rer cent 88-90 95-98 Filtrate quality Needs polishing by Already 0.2p

further filtrati on filtered

For the performance of the plant sca le system based on paramete rs g iven in Tabl e 2, TFF exercises for production size batches were conducted and data for five typical batches a re g iven in Tabl e 3.

The te tanu s toxin c larificati o n by depth filtration showed a recovery of 88-90 per cent in about 12-1 3 h, including post operation saniti zat ion. T he TFF sys te m, on the other hand , gave a better y ie ld of 95-98 per cent and required only 6.5 h for the filtration, inc luding the post-operation saniti zat ion . The system's performance, as compared to depth filtrati on is g iven in Table 4 .

The TFF system offered foll owi ng add ition al bene fits:

Closed operati ons, no spillage and thu s more in keep ing wi th GMP.

RAVETKAR eta!. : PROCESSI G OF TETAM US TO XI 777

Table 5-Details of processing of different production sca le batches using TFF system

Parameter Crude tox in lot No.

1467 1469 1471 1475 1477

Filtration date 29/03/97 09/04/97 19/04/97 I 0/05/97 20/05/97

Volume, L 1000 1000 1000 1000 1000

Permeate, Lf/mL 80 80 90 100 90

Saline used, L 60 60 60 60 60

Filterate, L 1040 1020 1025 1020 1035

Toxidation completion date 28/04/97 09/05/97 19/05/97 09/06/97 19/06/97

Concentration date 19/05/97 31/05/97 09/06/97 01/07/97 11/07/97

Concentration volume, L 42 41 40 52 42

Concentrate, Lf/mL 1900 1900 2000 2000 2000

Concentrate recovery, per cent 96 95 87 100 90

Salting out date 20/05/97 02/06/97 11/06/97 03/07/97 13/07/97

Final filtration date 23/05/97 04/06/97 13/06/97 05/07/97 16/07/97

Final filt rate volume, L 19.7 20.0 20.0 20.0 19.8

Filtrate, Lf/mL 3200 3200 3300 4250 3500

Fractionation recovery, per cent 79 82 83 82 83

Purity, Lf/mg PN 1957 1852 1670 1782 1852 l

Total recovery, per cent 76 78 72 83 74

Crude toxin lot No.

1479 1481 1482 1484 1488

Filtration date 30/5/97 10/06/97 20/06/97 30/06/97 21/07/97

Volume, L 1000 1000 1000 1000 1000

Permeate, Lti'ml 100 100 100 80 90

Saline used, L 60 60 60 60 80

Filtered volume, L 1035 1040 1020 1030 1050

Toxidation completion da~ 29/06/97 I 0/07/97 20/07/97 30/07/97 20/08/97

Concentrated on, date 2 1/07/97 0 1/08/97 II /08/97 23/08/97 I 0/09/97

Concentrate volume, L 47 50 52 37 43

Permeate in concentrate, Lf/mL 2000 1900 1900 1900 1800

Concentration recovery, per cent 9 1 91 97 85 82

Salting out date 23/07/97 03/08/97 13/08/97 25/08/97 14/09/97

Final filtration date 26/07/97 06/08/97 17/08/97 27/08/97 17/09/97

Final filtrate volume, L 20.0 20.0 20.0 20.0 20.0

Final filtrate , Lf/mL 3900 3750 4200 2750 3000

Fractionation recovery, per cent 83 76 85 78 78

Purity, Lti'mg PN 1800 1720 1929 1679 1542 l

778 J SCI INO RES VOL 60 OCTOBER 200 1

Minimal exposure of the working personnel.

No recurring cost of pads, etc.

Savings on autocla vin g charges for sterili zation of filter press assembli es.

Sav in gs on disp osa l ex pe nses lik e decontamination, and incinerat ion of used pads .

o problem of final sterili zation- filtration. as 0.2 Jl filtered output was used.

Validation of TFF System

The pre- and post-usage integrity tes tin g showed repeatable use of membrane device and fulfillment of an important GMP norm. For removal of formalin , the system must be flu shed with I 00 L of di sti lied water whereby the concentration of formalin is reduced below 5 ppm.

Absence of Cl tetani colony on the GVWP disc confirmed that no organi sm was present in the clarified broth . The positive growth of the organism on M & M agar also confirmed that the medium could support the growth of Cl tetani .

In the case of removal of acetic acid from the system, the test sample obtained after pass ing 50 L di stilled water had colour similar to that of with I 0 ppm acet ic acid after add ition of methyl red indicator. The colour change below thi s concentration was not di stin guishable. To ensure that no trace of the chemical remained, another 50 L distilled water was flush ed through the system.

In the case of the remo va l of sodium hypochl orite from the system, the test sample obtained after pass ing 50 L di stilled water had colour similar to I 0 ppm sod ium hypochl orite after addition of the uni versa l indicator. The colour change be low thi s concentrati on was not

distinguishable, and to further ensure that no trace of the chemical remained, another 50 L di stilled water was flush ed through the system.

Thu s, from GMP point of view th e system was completel y va lidated .

Concentration and Ammonium Sulphate Fractionation of Toxoid Using Pellicon System

As desc rib ed above , afte r co mpl e ti o n of detox ificati on, the toxoid was concentrated, ammonium sulphate fracti onated and fin all y sterile fil te red. The steps involved , suc h as c la rifi ca ti o n. co nce ntrati on, diafiltration and sterile filtration , in the production of tetanus toxoid are summari zed in Table 5.

The advantages of using a TFF system have been di scussed above. The benefits of Pellicon system used for diafiltrati on are: (i) It reduces the time required in dialysis, (ii ) The system can be va lidated as per the GMP requirement , and (iii ) There is minimum ex posure of working personnel, as it is a closed system.

Acknowledgments

We are ex tremely thankful to Dr C S Poonawalla. Pres ident and Dr J M Mehta , vice-president, Serum Institute of India Research Foundati on for their kind help during the research on thi s project.

References

Ran U B. M ahadevan M S & Michae ls S. Phmw 7i,cluwl. ( I 902 ) 102.

2 Levine, H.L. & Castillo, F.J , In 1/iotechnologr. qua/it\' as.wrance and validation , Drug Manufactcring Technology Series. ed ited by K E Avis. C M Wagner and V L Wu ( Inter Pharma Press Inc. Bullalo Grove. Illinois) Vol. 4, 1999 , I 54.

3 WHO, Manualfor the flroduction & contm l of' vaccine: TetWIIIS

toxoid (World llea lth Org~nizationlBLG/U DP/77.2 Rev. I