As seen from the previous graphs, exactly the same separation was obtained on both the development 10 ml axial column and the production 20 L radial flow column. Thus showing a process can be developed on AFC and transferred to RFC in production. Also what should be noted is that higher binding of protein per ml of resin was observed using the radial flow column as pressure did not build up as in the case with the axial column. Another point to note is no process optimization was done when the process was scaled up.

Conclusions:

This study shows an easy scale up from a laboratory protocol by a simple extrapolation of the conditions determined at a smaller scale. What is interesting to note is that besides obtaining higher flow rates resulting in a fast separation, the radial flow column also achieves a higher purification factor and activity yield leading to the conclusion that they provide higher productivity and a more efficient separation as compared to the axial column. This also shows that work done on an axial flow can be simply linearly scaled up to a radial flow configuration.

Conclusions:

Monoclonal Antibodies (MAB) are increasingly being prepared in large quantity from cell culture. Conventional purification methodologies are generally slow and not easily scaleable from milligram to multigram levels. Experimental conditions such as flow rates and gradient rates cannot directly be extrapolated from one level to the next. Reoptimization of flow and elution conditions can lead to lost time and more importantly, loss of valuable product for such empirical studies. Superflo® Columns have a patented radial flow design that allows fast separation even with conventional soft media. The unique design of these columns easily allows the user to predictably scale up the separation in a linear fashion. The flow rates and gradient volumes can thus be increased in proportion to sample size without altering the elution profile or purification time.

The desalted crude extract was fractionated on a 5 X 10 cm Pharmacia K-100 column packed with the highly cross linked resin DEAE-Sepharose®* fast flow equilibrated with the above buffer. The sample was loaded and the column ran at a flow rate of 12 ml/min. Fractionation was achieved with a linear 500 ml gradient of 0 to 0.75 NaCl with the above buffer. Enzyme activity eluted in a single peak. After sample loading, the time required to perform gradient elution was 60 minutes and a total run time of 175 minutes.

For the Superflo® 1200 column, 1.3 Kg of a bacterial
cell paste was thawed, washed, lysed and cell debris was removed in essentially an identical manner as described above, except that the sample was several fold more concentrated having a total volume of 1.88 L. Desalting was performed as described above prior to fractionation.

Fractionation was performed on a Superflo® 1200 packed with DEAE-Sepharose®* fast flow equilibrated with the above buffer and operated at a flow rate of 250 ml/min. Elution of enzymatic activity occurred in a single sharp peak with a linear 5.0 L gradient of the above elution buffer. After sample loading, the time required to perform gradient elution was 24 minutes and total operation time was 80 minutes.

No noticeable backpressure was observed with either column. In a subsequent column, run flow rates of 350 ml/min with no significant pressure and essentially identical elution profile were achieved. This suggests higher flow rates are possible with no significant loss in performance.

Differences in sharpness of peaks are probably not significant because of the precision of the assay.

Summary:

As seen from the chromatograms, exactly the same separation was obtained on both the development 200ml Axial Column and the production 1200 ml Radial Flow Column. Thus, a process can be developed on AFC and transferred to RFC in production. What should also be noted is that a higher binding of protein per ml of resin was observed using the Radial Flow Column. The flow rates were higher in the Radial Flow Column with no pressure build up, which happened in the axial column. This data suggests that processes can be converted from Axial columns to Radial Flow Columns with no optimization required. Bed height does not seem to play a role in this kind of on-off separation.

Conclusions: