Plasmodium falciparum W2 (MRA-157) was cultured using previously described methods [⁸]. Briefly, parasites were grown in human erythrocytes in RPMI 1640 media supplemented with 0.25 % Albumax II (GIBCO, Life Technologies, San Diego, California, United States), 2 g/L sodium bicarbonate, 0.1 mM hypoxanthine, 25 mM HEPES (pH 7.4), and 50 ?g/L gentamycin, at 37?C, 5% O₂, and 6% CO₂. Parasites were loosely synchronized with a single round of sorbitol treatment prior to magnetic purification.

The following neodymium rare-earth magnets were purchased from Applied Magnets (Plano, TX): paired 1.5 in OD ? 1.25 in ID ? 0.75 in long ? 90 degree grade N42 arcs (NA005), a diametrically magnetized grade N42 1 in diameter cylinder with a 0.25 in center hole (ND062-D), diametrically magnetized grade N42 1 in diameter cylinders (ND062-TD), and grade N52 1 in cubes (NB041). Magnetic fields were measured using a DC magnetometer (AlphaLab Inc., Salt Lake City, UT).

Synchronized cultures at 2% haematocrit were grown to a ring parasitaemia of 10% followed by two additional days of culture to allow natural egress and release of haemozoin. Each LS column (Miltenyi Bioec, Auburn, CA) was loaded with 25 ml of resuspended culture, washed three times with 5 ml water, and eluted in 5 ml water. Haemozoin was recovered by centrifugation for 15 min at 12,500 ? g and resuspended in 100 ul water. The final concentration of the recovered haemozoin was determined by dissolution into monomers and spectroscopic determination at 400 nm using a molar extinction coefficient of 10⁵, as previously described [⁹].

Cultures were synchronized by one round of sorbitol synchronization and purified once the parasites reached the late-trophozoite stage. Each LD column (Miltenyi Biotec) was loaded with 50 ml of parasite culture at 2% haematocrit and 5-15% parasitaemia. Columns were washed once with 5 ml complete RPMI medium (RPMIc) and eluted in 5 ml RPMIc. Recovered cells were counted on a haemocytometer and purity was assessed by Giemsa-stained blood smear.

The purification stand was designed in SolidWorks 2009 (Dassault Syst?mes SolidWorks Corp., Concord, MA) as three separate components (body, lid, and legs). The body accommodates five 1 in diameter ? 1 in length diametrically magnetized cylinders. Designs were exported as STL files and prototyped in acrylonitrile butadiene styrene (ABS+) plastic on a uPrint 3D printer (Dimension Inc., Eden Prairie, MN). The magnetic stand design files, appropriate for 3D printing, are available as supplementary files (Additional File 1). Updated versions are available in the 3D design repository at the DeRisi lab website [¹⁰].

Percentages were arcsine transformed as summarized by Zar [¹¹]. Means and confidence intervals were calculated on the arcsine-transformed data and converted back to linear space. All reported p-values are calculated from two-tailed t-tests assuming unequal variances on the arcsine-transformed data.

Naturally generated haemozoin has traditionally been prepared through a series of extractions to remove all traces of contaminating nucleic acids, proteins, and lipids [¹²]. More recent studies examining the immuno-stimulatory properties of haemozoin have turned to the use of magnetic purification in order to isolate haemozoin in a state closer to its physiological form upon release from parasites undergoing egress from erythrocytes [⁷]. Although magnetic purification is increasingly employed in the field of malaria research, the relationship between magnetic field strength and haemozoin yield during magnetic purification procedures has not previously been examined.

Therefore, the performance of three configurations of permanent rare-earth neodymium magnets producing different strength fields was compared with a commercially available MidiMACS separator (Miltenyi Biotec). Purifications were performed on LS columns, which have a relatively fast flow rate that allows retention of haemozoin, but not late-stage parasites, as previously described [⁷]. For the weakest magnetic field, four arc-shaped motor magnets were used to create a two-inch long channel with a oval-shaped cross-section, resulting in a magnetic field of approximately 2,000 G (measured at the estimated centre position of the purification column; Figure 1). The MidiMACS separator possessed the third strongest field, measuring approximately 4,500 G. The second strongest field was generated by a diametrically magnetized cylinder (1 inch long ? 1 inch diameter, with a ? inch hole down the centre), with a field of 5,100 G (estimated at a distance of one LS column radius from the surface). The strongest configuration was measured halfway between two one-inch cube magnets separated by a distance of one LS column diameter, measuring 8,400 G.

Haemozoin was purified in parallel over LS columns in each magnet configuration and analysed for yield. The relationship between the measured magnetic field strength and concentration of the recovered haemozoin was roughly linear, with the strongest magnetic fields resulting in the highest haemozoin yields (Figure 2). The apparent deviation from linearity exhibited by the cube magnets suggests that further increases in the strength of the magnetic field may result in diminishing returns. Most importantly, the neodymium magnets provided considerably higher yields than the MidiMACS separator (42% more for the cylinder, 78% more for the cubes), which is the current standard in the field.

A previous study suggested that "better separation" of iRBCs might be achieved through the use of strong electromagnets [³]. However, it has remained unclear how different magnetic fields might influence the purity, yield, and viability of isolated iRBCs, especially in comparison with commercially available systems that are commonly used for this purpose. A comparison of iRBC purification on slow flow-rate LD columns (Miltenyi Biotec) subjected to the magnetic fields from the MidiMACS, cylinder magnet, and double cube magnet configurations was, therefore, undertaken.

Blood smears of the recovered iRBCs demonstrated high purity for all magnets (Table 1). The order of overall iRBC yield was similar to that for haemozoin, with cubes giving the highest yield and the MidiMACS separator giving the lowest yield. Notably, the yield from both the cylinder and cubes was considerably better than that from the MidiMACS separator (39% and 42% more than MidiMACS, respectively).

Although the two-cube configuration provided the highest yields in haemozoin and iRBC purifications, the cube magnets are considerably more difficult and dangerous to work with than the cylinder magnets as a result of having a stronger pull-force and no convenient surfaces for grasping when dislodging them from a surface. For practical application, a multiple diametrically-magnetized cylinder (ND062-TD) configuration, which generates magnetic fields that are almost as strong (7,500 G) as the pair of cubes used in previous experiments, was employed. A four-column stand was designed using CAD software, and a fully functional prototype was constructed on a 3D plastic printer (Figure 3). The design includes four column positions interdigitated with five cylinder magnets. Because the cylinders are able to rotate within their chambers, they automatically orient themselves when juxtaposed to adjacent magnets, which also maximizes the strength of the magnetic field at the column positions. The total cost of the system was approximately $40 for the cylinder magnets and $35 for the stand. The most recent versions of the CAD design files for these parts are freely available at the DeRisi lab 3D design repository [¹⁰].

Late stage parasites from synchronous cultures of P. falciparum were isolated in parallel using both the in-house-designed purification stand and three MidiMACS separators. The recovered parasites were assessed for overall yield using a haemocytometer, purity by blood smear, and viability by culture. The cylinder magnets recovered approximately 60% more parasites than the MidiMACS separators (25.9% and 15.0% total recoveries, respectively; p = 0.0002). Purity was high and comparable for both methods (cylinders 95.3% mean with 95% CI 93.6% to 96.9%; MidiMACS 95.7% mean with 95% CI 92.5% to 98.0%; p = 0.86).

In order to assess viability after being subjected to different strength magnetic fields, the purified parasites were diluted to 1% parasitaemia and monitored for growth over the next three days. The purified parasites from the two different preparations exhibited growth that was essentially identical, with a plateau between rounds of egress and reinvasion (Figure 4). These results indicate that the use of inexpensive high-strength neodymium magnets for malaria parasite purification results in significantly higher yields than commercial systems, with no adverse effects on parasite viability.