For petri dishes or flasks containing red cells in a nutrient medium a major threat is contamination by bacteria (despite gentamycin in the medium) or by fungi: therefore, a Biosafety cabinet (Class II) is the main piece of equipment required (see Fig. 2A, B). The cabinet is equipped with a HEPA (High Efficiency Particulate Air) filter, capable of retaining 0.3-micron particles with 99.99% efficiency. We made sure that the airflow met specifications and that the cabinet was regularly serviced. We installed a UV lamp to be permanently on when the cabinet was not in use. Then, with the UV lamp turned off, we exposed open blood-agar and nutrient agar microbiology plates for 12 hours and confirmed there was no bacterial growth. The cabinet was always kept free of any unnecessary items. All manipulations involving cultures or reagents needed for cultures were carried out in this cabinet with sterile precautions. Wearing of gowns and of gloves is recommended; sleeveless gowns are best.
A low oxygen (about 13%) gas phase is known to be essential for optimal growth of P falciparum. Rather than continuous flow of a gas mixture from an ad hoc cylinder, we chose the so-called ‘candle jar’ approach for several reasons. (i) It is free of charge. (ii) Supply of the appropriate gas mixture cylinders may be erratic. (iii) In a sealed candle jar, if it is sterile to begin with, the cultures are completely protected from contamination (the same is not true in an incubator with continuous gas flow). The jar we used was a vacuum desiccator made of heavy glass (see Fig. 2C) with a 2 cm ground glass edge, and the lid has a similar edge (we found vacuum desiccators made of plastic not reliable). In order to obtain a perfect seal, we apply a thin but generous layer of high vacuum grease (Dow Corning Corporation, USA) to both edges and to the ground glass device incorporating the tap. The jar, when open, is handled only under the biosafety cabinet. We lay the flasks or dishes inside the jar, light a white candle, and put in place the lid with the tap open; when the flame goes out, we immediately close the tap.
The sealed jar is then carefully transferred to the incubator, that must have a good temperature control, and must be checked to be never outside the range of 36.8-37.1°C (see Fig 2D).
Culture of established P falciparum strains
Thanks to the courtesy of many colleagues (see Table 1) we have obtained several culture samples, some live and some frozen. The data in Table 1 indicate that, despite our precautions, infection was a significant problem especially at the beginning. In some cases, cultures may have failed because frozen parasites were no longer viable as a result of prolonged storage or problems associated with transportation.
Cultures of P falciparum from clinical isolates
In our attempts to culture parasites from patients we have selected, for obvious reasons, those who had high parasitemia (Table 2). In 9 attempts (leaving aside one in which the culture suffered early bacterial contamination), we saw production of new rings (Fig. 3A) in 8 cases: however, in 4 of these parasite growths stopped after one to five cycles. In the remaining 4 cases we obtained continuous cultures, but two of these were later lost (again because of bacterial contamination). With PAT-2 and PAT-6 we were able to document protracted exponential growth (Fig. 4). The multiplication factor per cycle (48 hrs) of clinical isolates ranged from 1.6 to 5.5, whereas it was 8.0-11.1 for the NF54 strain.
Composition of culture media
Since the original notion of Trager & Jensen that a strong buffer (HEPES) was required, and that 10-20% human serum would help to optimize growth, attempts to improve culture media have not gone far: except that human serum has been often replaced by bovine albumin (Albumax). For a start we preferred human serum because it is easily available and free of charge from generous donors; however, we were aware that human serum in a malaria-endemic setting is likely to contain antibodies that may inhibit P falciparum growth. In several experiments we observed that a 1:1 mixture of human serum with Albumax was either equivalent or superior to Albumax alone (see Fig. 5).
We have used the sorbitol technique (19) and the refrigeration technique (20). Starting from a culture with a parasite count of 8.6%, of which 65% rings, 25% trophozoites and 20% schizonts, we obtained a culture that had 77% rings after one round of sorbitol treatment, and 92% rings after two rounds of sorbitol.
Recovery of frozen parasites
Ideally parasitized red cells should be stored frozen in liquid nitrogen (i.e. at -195oC). However, since this was not available, we have stored parasitized red cells in a -80oC deep-freezer and recovered them successfully after up to 120 days.