1

In vivo splenic clearance corresponds with in vitro 1

deformability of red blood cells from Plasmodium yoelii 2

infected mice 3

4

Sha Huang1,2, Anburaj Amaladoss3, Min Liu3, Huichao Chen4, Rou Zhang3, 5

Peter Preiser3, 6, Ming Dao3,5*, Jongyoon Han1,2,7,* 6

1 Department of Electrical Engineering and Computer Science, 7 Massachusetts Institute of Technology, 8

77 Massachusetts Avenue, Cambridge, MA 02139 9 10

2Biosystems and Micromechanics (BioSyM) IRG, 11 Singapore-MIT Alliance for Research and Technology (SMART) 12

1 CREATE Way, Singapore, 138602 13 14

3 Infectious Diseases IRG, 15 Singapore-MIT Alliance for Research and Technology (SMART). 16

1 CREATE Way, Singapore 138602 17 18

4 Department of Biostatistics, 19 Harvard School of Public Health, 20

651 Huntington Ave, Boston, MA 02115 21 22

5 Department of Materials Science and Engineering, 23 Massachusetts Institute of Technology, 24

77 Massachusetts Avenue, Cambridge, MA 02139 25 26

6 School of Biological Sciences, 27 Nanyang Technological University, 28

60 Nanyang Drive, Singapore 637551 29 30

7 Department of Biological Engineering, 31 Massachusetts Institute of Technology, 32

77 Massachusetts Avenue, Cambridge, MA 02139 33 34 35

*Corresponding Authors: 36

Jongyoon Han ( jyhan@mit.edu ) 37

Tel: (617)-253-2290; Fax: (617)-258-5846 38

or 39

Ming Dao (mingdao@mit.edu) 40

Tel: (617)-253-2100; Fax: (617)-258-0390 41

IAI Accepts, published online ahead of print on 31 March 2014Infect. Immun. doi:10.1128/IAI.01525-13Copyright © 2014, American Society for Microbiology. All Rights Reserved.

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

2

Abstract 42

Recent experimental and clinical studies suggest a crucial role of mechanical 43

splenic filtration in the host’s defense against malaria parasites. Subtle changes in red 44

blood cell (RBC) deformability, caused by infection or drug treatment, could influence 45

the pathophysiological outcome. However, in vitro deformability measurements have 46

not been directly linked in vivo with the splenic clearance of RBCs. In this paper, mice 47

infected with malaria-inducing Plasmodium yoelii revealed that chloroquine treatment 48

could lead to significant alterations to RBC deformability and increase clearance of both 49

infected and uninfected RBCs in vivo. These results have clear implications for the 50

mechanism of human malarial anemia, a severe pathological condition affecting malaria 51

patients. 52

53

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

3

INTRODUCTION 54

The spleen serves as the largest filter of blood in the human body by initiating 55

immune responses against blood-borne microorganisms and removing abnormal blood 56

cells (1). Two major structural features of the spleen enable its critical functions: the 57

white pulp (including marginal zone), which contains the majority of immune effector 58

cells, and the red pulp, a reticular meshwork that filters abnormal red blood cells (RBCs) 59

(2). In humans, 76-79% of the spleen is made of red pulp, a dense meshwork composed 60

of splenic cords and splenic sinuses. Previous studies on different animal (dog, cat and 61

rat) models reveal that from 90% (3-5) to 10% (6) of total splenic blood undergoes so-62

called “closed circulation”, during which RBCs traverse venous sinuses bypassing the red 63

pulp; open circulation occurs in the remaining blood whereby RBCs enter the reticular 64

meshwork, following slow microcirculation (1, 7). The structural and mechanical quality 65

of the RBCs is ascertained by the mechanical constraint imposed by the meshwork in the 66

red pulp where old and abnormal RBCs that are less deformable are retained and 67

eventually removed by phagocytosis (7). 68

The important implications of RBC deformability in the pathogenesis of malaria 69

have been extensively discussed (8-13). Significant decrease in RBC deformability arising 70

from Plasmodium falciparum invasion was observed in vitro with different 71

measurement techniques including optical tweezers (9), micropipette aspiration (13), 72

RBC membrane fluctuations (i.e. diffraction phase microscopy) (12) as well as the 73

probing of cells under flow (i.e. microfluidic flow cytometer) (10). Such parasitization of 74

RBCs has been shown to increase their stiffness many-fold, with their elimination by 75

mechanical filtration expected to compromise their microcirculation (8). 76

The connection between RBC deformability and splenic clearance has been 77

demonstrated in a series of ex vivo spleen studies (7, 14-17). Splenic retention of both 78

ring-stage malaria infected RBCs (iRBCs) as well as artificially hardened (17) (by heating) 79

uninfected RBCs (uRBCs) was observed via ex vivo perfusion of human spleen (14). It is 80

evident that, besides possible molecular interactions, the mechanical properties of RBCs 81

play a vital role in the process of splenic RBC clearance. This was further validated by 82

experiments that mimicked splenic retention in vitro using a microsphere filtration 83

system (16). 84

In fact, the role of the spleen in influencing the pathogenesis of malaria has been 85

well documented in a number of clinical studies. Splenomegaly (enlarged spleen) is a 86

characteristic clinical consequence of malaria infection, and therefore, the size of spleen 87

has in fact been used to estimate the intensity of malaria transmission (2). Clinical 88

studies involving radioactively labeled RBCs reveal that patients with an enlarged spleen 89

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

4

display a more rapid clearance of RBCs compared to patients with a normal spleen (18). 90

It has been proposed that splenomegaly modifies blood microcirculation and splenic 91

filterability (2). Studies on splenectomized hosts that show higher fatality rates and 92

delayed parasite clearance after antimalarial treatment (19) also point to the role of the 93

spleen in the clinical outcomes for malaria patients. 94

Experiments also suggest that splenic retention of RBCs could contribute to 95

malarial anemia (17), which is a common consequence of severe malaria associated with 96

high mortality (20). Excessive splenic clearance of RBCs is considered a likely mechanism 97

for malarial anemia (17). However, removal of only the iRBCs cannot be the primary 98

cause of such massive blood loss (20), particularly since clinical studies do not find a 99

correlation between severe malarial anemia and a high parasitemia level in the patient 100

(21, 22). These observations suggest the possibility that excessive clearance of uRBCs 101

could play a key role in the development of malarial anemia. While this process is not 102

fully understood, several mechanisms have been proposed for the increased clearance 103

of uRBCs, including the activation of splenic macrophages and enhanced splenic 104

mechanical retention by altering the mesh size of spleen red pulp (20). 105

How antimalarials influence RBC retention in the spleen could consequently 106

impact on the therapeutic outcomes of malaria; however the mechanical impact of 107

chloroquine on host RBCs has not been explored. In the past, the inhibition of hemozoin 108

formation was inferred as a key mechanistic consequence of chloroquine treatment (23). 109

Hemozoin is a non-toxic crystal synthetized by the parasites as they digest RBC 110

hemoglobin and release highly toxic (ferriprotonporphyrin IX) α-hematin (24). Since 111

hematin may lead to RBC membrane disruption and eventually host cell lysis (25), the 112

parasites need to convert hematin to hemozoin for their own survival. It is known that 113

chloroquine can prevent hematin polymerization (23), but whether it also modifies host 114

RBC deformability and consequently alters splenic RBC retention is unknown. 115

Although ex vivo studies (14) of human spleen show retention of both iRBCs and 116

artificially hardened uRBCs, little is known about the RBC splenic clearance during the 117

course of malaria infection and antimalarial treatment. Furthermore, a better 118

understanding from in vivo studies of possible connections between the mechanical 119

retention of RBCs in the spleen and excessive blood loss or anemia is needed. The 120

dynamic splenic response interweaves host, parasite and antimalarial drugs in such a 121

complex manner that ex vivo spleen studies alone are insufficient to predict the role of 122

splenic retention in influencing any anemic response in the host. Therefore, a 123

quantitative and more direct measurement of the deformability of both spleen minced 124

blood (SMB) (more details in animal preparation under Materials and Methods section) 125

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

5

and peripheral venous blood (PVB) in healthy and malaria infected host is highly 126

desirable. 127

Rodent malarial model has been commonly used to complement the research on 128

Plasmodium falciparum (20, 26). Splenic clearance of parasitized RBCs was determined 129

to play an important role in both human (17) and mice (27). In this paper, we identified 130

Plasmodium yoelii as the most relevant rodent model to study in vivo splenic RBC 131

clearance for it shares similar invasion characteristics (28) with Plasmodium falciparum. 132

Careful consideration was also given to the structural similarities and differences 133

between human and mice spleens: human spleen is sinusal (29); human RBCs (8 µm) 134

have to squeeze through the interendothlial slits (~1 µm) (16) in venous sinus walls 135

which act as a mechanical filter to abnormal or stiffened RBCs (5). In comparison, 136

though mouse spleen is arguably classified as nonsinusal (5), the fenestrations in the 137

walls of mouse pulp venules are so small (1-3µm) (30) compared to murine RBCs (~6µm) 138

that they still function just like venous sinus and mechanically trap less deformable RBCs 139

(31). 140

To assess the deformability of RBCs in mouse spleen, we extracted mouse RBCs 141

from both PVB and SMB and the deformability of the RBCs was quantitatively evaluated 142

using a microfluidic deformability cytometer (10). Several important aspects relating to 143

splenic RBC retention were explored: first, we established the correlation between in 144

vitro deformability assay and in vivo splenic retention; secondly, the effects of malaria 145

infection and/or antimalarial drug treatment on RBC deformability as well as on splenic 146

RBC retention were then investigated; thirdly, we attempted to use 2 different 147

approaches to estimate splenic retention threshold based on RBC deformability. 148

Physically, retention threshold is decided by the effective pore size of the reticular 149

meshwork or the splenic slits, as well as the RBC geometry (size, shape) and membrane 150

stiffness. In this study, we estimate retention threshold below which RBCs are 151

considered to be most likely retained, expressed as a fraction of normalized PVB velocity. 152

Finally, the possible anemic effect related to increased splenic retention, both in 153

infected and uninfected mice and with or without drug treatment, was investigated. 154

155

156

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

6

MATERIALS AND METHODS 157

Murine model for malaria-infection 158

4 to 6 weeks male or female Balb/C mice were infected with 1x105 parasites of 159

Plasmodium yoelii YM by i.p. injection. This blood smear made from these infected mice 160

showed 1-10% parasitemia approximately 4 days post-infection. Mice were then 161

injected with drug or PBS for 3 consecutive days by i.p. injection as described below. 162

Microfluidic deformability measurement 163

The deformability of single RBC was assessed using microfluidic deformability 164

cytometer (Figure S1 B) as described (10). Blood samples were diluted to 0.1 – 1% 165

hematocrit before loading to the system to minimize cell-cell interactions. The system 166

was operated at pressure driven mode. RBC movement in the main channel was 167

captured by CCD camera and the video could be post-analyzed using ImageJ software. 168

The deformability for every RBC was characterized by its average traverse velocity 169

across repeated bottleneck structures. The device channel height is 4.2µm which 170

ensures white blood cells and other cells from splenic minced blood to stay in the 171

reservoir and not enter the main channel. However, we note that due to the extremely 172

complex cell mixtures in spleen minced sample (especially from the infected mice), 173

device throughput is limited. In the present study, we still managed to measure a 174

reasonable sample size, which are larger than other conventional RBC deformability 175

measurement methods such as micropipette aspiration or optical tweezers, for all 176

conditions. 177

Animal preparation 178

Adult BALB/c mice each weighing approximately 20g were used for our 179

experiments. Four treatment conditions were included: healthy mice/saline, healthy 180

mice/CQ, malarial-infected mice/saline, and malaria-infected mice/CQ. Prior to the 181

experiments, approximately 10 µl venous blood was taken out from all mice for baseline 182

measurements of both RBC deformability studies and Hb assays. During the experiments, 183

the healthy mice were bled approximately 10 µl on alternative days and the malarial-184

infected mice were bled only on the first and the last experimental days. The spleens of 185

all mice were harvested when they were culled. Spleens were minced with sterile 186

scissors and forceps. The extracted blood was washed three times before being loaded 187

to device. 188

Drug treatment 189

Chloroquine diphosphate salt (Sigma-Aldrich) was dissolved in DI water with a 190

final concentration of 100mM, stored at -20 oC freezer. Working dose was then 191

prepared weekly by diluting the stock solution with 1X PBS buffer at the ratio of 2:11, 192

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

7

stored at 4 oC. The mice were received 100µl diluted CQ solution daily via IP injection. In 193

the control groups, mice were injected with 100 µl sterile PBS buffer solution. 3 194

consecutive days of CQ/ or saline treatments were carried out for all malaria infected 195

mice and their spleens were harvested. For healthy mice group, drug effect was studied 196

over CQ treatment days. Mice received 4, 6, or 8 days of consecutive CQ/ or saline 197

injection before being culled on the following day (Figure S1 A). 198

Sample preparation 199

To differentiate iRBCs from uRBCs, Hoechst dye (33342, sigma) was added to the 200

sample 20 minute prior to the microfluidic experiment so that the infected cells were 201

fluorescent under UV excitation. A fixed inter-pillar gap size of 3µm was used for all 202

mice RBC deformability measurements to achieve best deformation differentiation. The 203

microfluidic device is pre-coated with 1% BSA (or 20% FBS) to minimize confounding 204

effects due to RBC adhesion. 205

Experimental flow chart 206

The deformability of red blood cells was investigated in uninfected or P. yoelii 207

(YM) infected balb/c mice. Figure S1A describes the flow chart of one single 208

experimental round, consisting of 6 healthy mice and 6 P. yoelii infected mice. At least 3 209

experimental rounds were performed to achieve sufficient mice and cell count for all 210

characterizations (including microfluidic deformability, microscopy morphology, 211

micropipette aspiration, hemoglobin assay, spleen mass and size quantification). For the 212

experiments involving only healthy mice, three of them received 750 µg (i.e. 213

approximately 30mg/kg) of chloroquine (CQ) drug via intraperitoneal (i.p) injection daily 214

(32, 33) and were therefore referred as Healthy-Drug mice. The remaining three healthy 215

mice received equal volume of 1X phosphate buffered saline (PBS) solution and were 216

termed as Healthy-Control mice. On experimental day 1, approximately 10µl of blood 217

was extracted from all healthy mice to establish baseline measurement. After 4 218

consecutive days of CQ or PBS treatment, one mouse of Health-Control and Healthy-219

Drug group each was culled on day 5. Both peripheral venous blood (PVB) and splenic 220

minced blood (SMB) (34) were collected for subsequent measurement. CQ and PBS 221

treatment continued for the remaining four healthy mice and the procedure was 222

repeated on day 7 and 9 until all mice were culled. For the experiments involving 223

infected mice, 105 parasites of P. yoelii YM were injected in all 6 mice 2-3 days prior to 224

the experiment. The parasitemia was monitored by Giemsa stained thin blood smear 225

every 24 hours. When parasite levels first reached > 1% for all infected mice, 10µl blood 226

was taken for baseline measurement (i.e. experimental day 1). At this stage three 227

infected mice received CQ drug treatment (the Malaria-Drug group), and the remaining 228

three mice received a PBS placebo (the Malaria-Control group) for 3 consecutive days. 229

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

8

All mice were culled on day 4, and the PVB and SMB were collected. At this stage the 230

Malaria-Drug mice had an average parasitemia less than 1% while the Malaria-Control 231

mice had an average parasitemia between 50 - 90%. It is noted that more than three 232

experimental rounds were carried out to ensure data reproducibility among different 233

animals and to achieve sufficient data points for statistical analysis. 234

Statistical testing 235

Two tailed Mann-Whitney test was used for p-value computations unless 236

otherwise specified. 237

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

9

RESULTS 238

Splenic RBC retention based on the RBC deformability profiles. RBC 239

deformability profiles were characterized by a cell deformability cytometer, a 240

microfluidic device which consists of triangular pillar arrays as described by Bow et al. 241

(10) (Figure S1B). The inter-pillar gap size was selected to be 3 µm for all mice RBC 242

measurements. Individual RBCs deformed considerably as they passed through the 243

device and their traverse velocities were recorded to describe a population-wide 244

deformability profile (Figure 1A&B). It is understood that higher traverse velocity 245

corresponds to higher deformability (10, 11, 13). The deformability (velocity) profiles of 246

RBCs sampled from PVB were compared to RBCs from SMB of the same mice. As the red 247

pulp of the spleen is capable of retaining the abnormal and hardened RBCs, a fraction of 248

splenic minced RBCs were attributable to the splenic trapped RBCs, while the remainder 249

would represent healthy normally circulating RBCs (34). Therefore, two significant RBC 250

subpopulations in the spleen minced RBCs were assumed: 1) “flow through” RBCs that 251

share very similar mechanical properties to the peripheral RBCs; 2) abnormal or 252

senescent RBCs that were trapped in the splenic meshwork (16, 35). Mathematically 253

the population wide splenic RBC velocity profile can be expressed as follows: 254 1 1

where , , and represent the average deformability (or velocity) of 255

SMB, flow through, and splenic trapped RBCs; denotes the fraction of SMB that came 256

from spleen-trapped cells. Comparison of the normalized SMB velocity against the 257

average velocity of the PVB sample of the same healthy mouse (Figure 1A) showed that 258

RBC velocity of SMB was on average 15% lower than that of PVB (p<0.001), reflected as 259

a very gentle left shift in the velocity histogram (Figure 1B). This difference in blood 260

deformability profiles was likely contributed by the second subpopulation in SMB 261

samples, reflecting the fraction ( . . of less deformable RBCs trapped in the splenic 262

meshwork. No significant difference in the size distributions of PVB and SMB RBC 263

populations was identified (Figure S7) 264

Similar analysis was also performed on P. yoelii infected mice and the infected 265

RBCs (iRBCs) were differentiated from uninfected RBCs (uRBCs) by Hoechst staining (36). 266

Figure 1C illustrated different deformability profiles of both uRBCs and iRBCs extracted 267

from SMB and PVB samples. The average uRBC and iRBC velocities of SMB were 268

respectively 21% and 30% slower than respective uRBCs and iRBCs in the PVB (p<0.001). 269

Splenic sequestration of less deformable RBCs is one likely explanation (15). The 270

reduced splenic iRBC velocity reconciled with increased membrane shear modulus 271

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

10

analyzed by micropipette aspiration (Table S10). The more pronounced difference 272

between PVB and SMB in the infected mice (21% and 30% as compared to 10% in 273

healthy mice) indicates a possibly increased splenic retention in mice after P. yoelii 274

infection. This was further supported when examining the size of the healthy and P. 275

yoelii infected mice spleens as it has been demonstrated clinically that splenomegaly is 276

linked with enhanced RBC clearance in spleen (17, 18, 37): the length of an infected 277

mouse spleen (Figure 1E) was approximately 1.6X of a healthy mouse spleen (i.e. ~4X 278

volume enlargement), suggesting a more intense RBC clearance is likely to have 279

occurred. The deformability of parasite-free RBCs from healthy (i.e. hRBCs) and infected 280

(i.e. uRBCs) mice as well as corresponding spleen sizes were summarized in 281

supplementary table S1C. 282

283

The influence of malaria infection or/and antimalarial drug on RBC 284

microcirculatory behavior and splenic RBC retention. There are clear differences in the 285

deformability of RBCs obtained from PVB from healthy and infected mice (Figure 2A & 286

2B). Normalized against average velocity of healthy RBCs (i.e. hRBCs) from healthy mice, 287

the average velocity of uRBCs from malaria infected mice was 0.91, 9% slower than 288

hRBCs (p<0.001), and the average velocity of iRBCs was only 0.58, 42% slower than 289

hRBCs (p<0.001). These results provide strong in vivo evidence that malaria parasite 290

infections not only significantly stiffen diseased RBCs, but also have a visible impact on 291

uninfected RBCs of the host. This direct impact on uRBCs could result in an increased 292

retention of these cells in addition to iRBCs in the spleen. 293

The mechanical impact of the antimalarial drug chloroquine (CQ) on infected 294

mice was investigated. In the peripheral blood, the average velocity of both uRBCs and 295

iRBCs dropped by 17% and 42%, respectively, after CQ treatment (p<0.01) (Figure 2C 296

and 2D). 297

The spleen harvested from an infected mouse after 3 days of CQ treatment was 298

21% longer and 53% heavier than the spleen of the other infected mouse receiving PBS 299

placebo (Figure 2G, 4B & Table S1C). The increased spleen size/weight is typically 300

associated with increased RBC as well as macrophage count (17, 37). In other words, 301

RBC congestion and retention in the red pulp could be an important contributing factor 302

for the observed splenomegaly. This observation was in good agreement with our 303

studies on splenic uRBC velocity (Figure 2E & 2F): the average velocity of splenic uRBCs 304

declined from 0.72 to 0.54 after CQ treatment (p<0.001) and the percentage of splenic 305

uRBCs moving at a velocity below 0.65 increased from 29% to 62% after CQ treatment, a 306

sharp contrast as compared to only 2.2% (19 out of 873) of untreated peripheral hRBCs 307

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

11

have velocity less than 0.65 (Figure S2). Details on the threshold value (0.65) 308

determination are discussed later. 309

310

Effect of antimalarial drug on healthy mice’s RBC deformability profiles in 311

peripheral blood and in spleen. Chloroquine was reported to be concentrated within 312

malaria parasites by the formation of hematin, explaining the selective drug toxicity (38). 313

On the other hand, high dose and/or prolonged chloroquine exposure showed 314

detrimental effects on the spleen and brain tissues of a healthy host: significant increase 315

in the protein and cholesterol level was found in mice with prolonged CQ exposure; 316

disorganization in red pulp was also reported (39). With our observed effect of CQ on 317

uRBCs from P. yoelii infected mice, it would be interesting to investigate the effect of CQ 318

on the deformability of hRBCs from healthy mice. 319

Healthy mice were treated with CQ for 8 consecutive days. For hRBCs sampled 320

from PVB, mean hRBC velocity dropped by 9% and 15% respectively after 4 and 6 days 321

of consecutive CQ treatment (p<0.001). No further velocity change was seen on day 9 322

(p>0.1, Figure 3A). For hRBCs sampled from SMB, in additional to a gradual decrease, 323

the velocity profile appears to assume bimodal distribution, in which two normal 324

distributions seem to start separate at normalized velocity close to 0.7 (Figure 3B). 325

Throughout the experiment, no significant size changes in the mice spleens were 326

observed. 327

328

Bimodal estimation of splenic hRBC velocity profiles after CQ treatment. To 329

study the distributions of hRBC velocity, Figure 3B was re-plotted as histograms (Figure 330

S3 A). We fit data with both bimodal density functions and single normal distributions 331

(Figure S3 B-E). A comparison of the two models via likelihood ratio test (LRT) suggested 332

that the bimodal distribution fit better to all SMB data (p < 0.05). 333

Since RBCs from SMB consist of two significant subpopulations of flow through 334

and trapped RBCs, the contribution of each subpopulation to overall velocity profile was 335

assessed using probability density estimation, where a1 denotes the fraction of cells 336

coming from the splenic retained population; b1, c1 denote the mean and standard 337

deviation of the spleen retained RBC velocities; and b2, c2 denote the mean and 338

standard deviation of the normal RBC velocities. Parameters were estimated by the 339

maximum likelihood (ML) method, and all fitted results were listed in supplementary 340

table S4. 341

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

12

∑ √ exp 2 342

where Φ(.) represents the cumulative distribution function of the standard normal 343

distribution. Based on the parametric fittings, (red curves in Figure S3), a1 was estimated 344

to be 0.39 in the control sample and increased to 0.68 after 6th days of consecutive CQ 345

treatment. CQ induced alteration in RBC microcirculatory behavior was hence suggested 346

by the increased fraction of splenic retained RBCs. Besides, except for the control 347

sample, the values of b1 and b2 remained largely unchanged, centered on 0.65 and 1.0 348

respectively. This result suggests a fairly constant splenic retention threshold. The value 349

of b2 agrees well with the average peripheral hRBC velocity of 1.0. 350

351

Malaria infection, antimalarial treatment, and blood hemoglobin concentration. 352

Malarial anemia is one of the most common complications of Plasmodium falciparum 353

malaria (20) and splenic RBC retention has been suggested to be a potential contributing 354

mechanism (17). The possible anemic effect relating to increased splenic retention was 355

hence investigated. 356

In mice infected with parasites the average hemoglobin level had dropped from 357

19.3 2.2 g/dl to 16.0 1.3 g/dl when the parasitemia first reached 1-10% (day 1, 358

Figure 4). Infected mice were then treated with either PBS placebo or CQ for three 359

consecutive days. By day 4, the parasitemia of Malaria-Control (i.e. PBS treated) mice 360

reached 50-90%; all displayed severe anemic syndrome with average hemoglobin 361

concentration of 3.8g/dl (day 4, Figure 4). In comparison, all Malarial-Drug (i.e. CQ 362

treated) mice had parasitemia well below 1%. Despite the very low parasite burden, the 363

average hemoglobin concentration of these Malarial-Drug mice was 10.3g/dl, still 364

significantly lower than that on day 1 before CQ treatment (16.0g/dl, p<0.01). It is also 365

noted that several Malarial-Control mice from different experimental batches died on 366

day 4 and were disregarded in all measurements. The death could be associated with 367

severe anemia and therefore extremely low hemoglobin concentration. The actual 368

hemoglobin concentrations in the Malarial-Control group were likely to be even lower. 369

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

13

DISCUSSION 370

While there have been several experimental results suggesting a significant role 371

of splenic RBC clearance in human malaria pathology, this has never been validated in 372

an in vivo setting with actual disease progression. This work aims to study in vivo 373

filtration of RBCs directly in a mouse malaria model, and connect it with in vitro 374

microfluidic based deformability measurement. Clinical studies have shown both 375

similarities and differences in the malaria pathology between human and mouse in 376

terms of invasion characteristics and malarial anemia (20). For the scope of our study, in 377

which mechanical RBC filtration spleen was concerned mainly, Plasmodium yoelii- 378

infected mice were chosen as our mouse model for two reasons: first, mouse spleen 379

trapping of Plasmodium yoelii-infected RBCs was demonstrated by prior experimental 380

studies (40); second, similar to Plasmodium falciparum infection in human that the 381

massive destruction of uninfected RBCs play a significant role in malarial anemia, non-382

parasitized RBCs in Plasmodium yoelii-infected mice are also believed to contribute to 383

murine malarial anemia (41). Additionally, it should be noted that for a typical 384

nonsinusal spleen, adhesion (rather than deformability) is believed as the predominant 385

mechanism for RBC removal, as the critical mesh size of the spleen is significantly larger. 386

However, mice spleen has very small fenestrations (1-3µm) that mechanically function 387

like venous sinus and RBC deformability therefore plays a major role in the splenic 388

retention in mice (30, 31). We would also like to point out that RBC surface properties 389

may indeed have an impact on our analysis, as well as on RBC clearance in vivo (42). To 390

minimize confounding effects from RBC adhesion, we pre-coated the microfluidic device 391

as well as RBC samples with 1% BSA or 20% FBS at least 20 min prior to the experiment. 392

It is worth mentioning that the deformability measured by a microfluidic device 393

may not be equivalent to other measurement of cell deformability, such as micropipette 394

and ektacytometry. Cell deformability is a complex characteristics, depending on many 395

factors such as shear rate that cannot be reduced to a single number (43). Still, our 396

deformability cytometer is arguably better than other static deformability 397

measurements, in terms of mimicking the splenic filtration process and predicting 398

individual cells’ filterability in vivo. Indeed, the percentages of most likely retained RBCs 399

as predicted using our velocity based single-cell measurements are in good agreement 400

with reported retention rates of Plasmodium falciparum RBC cultures (16) (S6). The 401

relatively low shear rate (<100 sec-1) and slow RBC flow velocity (up to 200 μm/s) used 402

in our system are consistent to those reported in vivo (44, 45) and the strong correlation 403

between in vitro deformability and in vivo splenic filtration profile, demonstrated in our 404

data, also support this argument. 405

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

14

RBC deformability, splenic RBC retention and malarial anemia 406

Severe malarial anemia frequently associates with high mortality rate in children 407

and pregnant women. RBC loss is considered as its primary reason. Several mechanisms 408

were proposed to account for such massive RBC loss. Antibody-mediated hemolysis, for 409

example, is believed to play a major role in patients with hyper-reactive malaria 410

splenomegaly. Such delayed hemolysis effect reflects diverse pathogenic processes and 411

intrinsically associates with RBC deformability both ways: on one hand, hemolysis is 412

partly caused by membrane defect and reduced RBC deformability, on the other hand, 413

hemolysis affects the deformability of surrounding RBCs (46). Only recently, splenic RBC 414

sequestration was proposed as another simpler mechanism for RBC loss (15): less 415

deformable RBCs get removed through the mechanical trapping of the spleen. 416

In the present study, a 37.5% drop in hemoglobin concentration (from 16.0g/dl 417

to 10.3g/dl) was found in the CQ-treated malarial mice, while these mice exhibited fairly 418

low parasitemia throughout. Such a big drop can’t be attributed mainly to the hemolysis 419

of iRBCs. On the other hand, the stiffening of uRBCs and the significant spleen 420

enlargement after CQ treatment suggest spleen related removal of uRBCs may play a 421

major role in observed anemia. The SMB displayed significantly lowered transit velocity 422

(Figure 1A, 1C, and 2C), which is in line with the spleen filtering less deformable RBCs. 423

These observations along with the earlier ex vivo results by Buffet et al. (14), establishes 424

that RBC deformability has an important contribution in in vivo splenic clearance. 425

Additionally, for the first time, we demonstrate the important links connecting 426

RBC deformability, splenic RBC retention and malarial anemia within the in vivo mouse 427

setting. Several earlier studies have eluded the possibility that intensified RBC retention 428

in spleen exacerbates malarial anemia (7, 15). Here we demonstrate that reductions in 429

RBC deformability and increased spleen size after malaria infection and CQ treatment 430

exhibit strong correlations with the hemoglobin concentration of the test subject. It 431

needs to be highlighted that CQ treatment, though successfully reduced the parasite 432

count in the infected mice, did not alleviate malarial anemia completely (Figure 4) 433

suggesting that instead of parasite loading alone, splenic retention of uRBCs could be 434

directly responsible for the excessive blood loss in malarial anemia. The significantly 435

lowered hemoglobin concentration (from 16.0g/dl to 10.3g/dl after CQ treatment) 436

appeared to correlate with the increased size and weight of mice spleens (Figure 4B), 437

which is consistent with reported clinical studies on human patients (17, 18, 22). After 438

all, it must be noted that our retention model is rather simplified. Splenomegaly is highly 439

multifactorial that other mechanisms such as stress erythropoiesis and immune 440

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

15

mediated RBC destruction would add in the understanding of the extremely 441

complicated and dynamic process. , 442

443

Chloroquine decreases RBC deformability and enhances splenic RBC retention 444

Another key finding of this study is that CQ has a direct impact on the red blood 445

cell deformability. This decrease in RBC deformability provides another mechanism of 446

action of CQ as it leads to increased retention of iRBC in the spleen. On the other hand 447

the fact that uRBCs show reduced deformability and consequently enhanced retention 448

in the spleen can explain the high level of anemia seen at low parasitemea levels. To the 449

best of our knowledge, we provide here the first demonstration of CQ induced iRBC 450

alteration on RBC deformability in vivo by measuring RBC velocities under flow. 451

The exact mechanism for CQ induced reduction in RBC deformability (or velocity) 452

is still unclear. Drug induced changes on RBC size and shape is one possible factor (47), 453

that could have direct impact on RBC deformability (35, 43, 48, 49). Indeed, significant 454

changes in the size and sphericity of iRBCs (trophozoites and schizonts) were observed 455

(Figure S9), and slight increase in stomatocytes appears to be evident after 5 days of 456

consecutive CQ treatment (Figure S8). 457

On the other hand, membrane stiffening can be another important source for 458

drug related iRBC rigidification as CQ causes hemin-induced oxidative damage on RBC 459

membrane (46). This possibility has been validated using micropipette aspiration; the 460

average membrane shear modulus of iRBCs (trophozoites and schizonts) was found to 461

have increased by 2-8 folds after CQ treatment (Table S10). The impact on uRBC and 462

hRBC in both healthy and malaria infected mice could be due to a general increase of 463

the oxidative stress in the system (50). An earlier study injecting chloroquine to healthy 464

swiss mice reported that, after 5mg/kg chloroquine i.p injection (approximately 8 fold 465

lower dose than our 0.8mg/mice) both Glucose-6-phosphate dehydrogenase (G6PDH) 466

and NADH diaphorase activities of normal RBC increased significantly (51), suggesting 467

that the oxidative stress induced by chloroquine needs to be compensated by increasing 468

the activity of protective enzymes. On the other hand, RBCs deficient in G6PDH are 469

more susceptible to oxidative damages, so that drug induced oxidative stress could 470

drastically decrease the deformability of these RBCs (52, 53). It is also possible that 471

chloroquine impacts RBC deformability through accelerating hemoglobin denaturation 472

(54). In all, the mild albeit significant reduction in hRBCs and uRBCs transit velocity (i.e. 473

deformability) after CQ treatment is not unique to choloroquine: in a separate work, we 474

also observed similar stiffening (in terms of both higher shear modulus and lower 475

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

16

microcirculatory velocity) on uRBCs in vitro, from Plasmodium falciparum malaria 476

infected human blood incubated with artesunate, another popular antimalarial drug 477

containing an endo-peroxide bridge (13). 478

479

Threshold prediction for splenic RBC retention 480

We are able to establish a clear relationship between RBC deformability and 481

splenic RBC retention in a more quantitative manner, enabling us to study the effective 482

threshold for retention. The term retention threshold was previously introduced (17) to 483

describe the stringency of spleen as a mechanical filter. It was postulated that different 484

retention thresholds may exist in healthy and malaria infected host. 485

The average lifespan of red blood cells (RBCs) in balb/c mice is approximately 40 486

days (55). Aged mouse RBCs, which are generally stiffer (56), are trapped by mice spleen 487

for phagocytosis (57). The spleen is therefore estimated to filter the least deformable 488

2.5% red blood cells during microcirculation each day (17). Based on this approximation, 489

a cutoff line was drawn on Figure 1D, marking the bottom 2.5% of PVB. The 490

corresponding velocity range allows a rough estimation of splenic retention threshold in 491

healthy animals, which is around 0.65 of normalized PVB velocity (Figure S2). 492

This threshold is cross-validated by the maximum likelihood parametric fitting 493

(S4), which resolved the SMB into two potential subpopulations according to the 494

velocity distributions. The average velocity of the spleen-retained subpopulation, as 495

suggested by parametric fitting, is also around 0.65 (i.e. , see supplementary table S4). 496

It is interesting to note that a shift in the retention threshold was observed in 497

mice after malaria infection. In contrary to healthy mice, of which the threshold was 498

estimated to be centered at 0.65 (i.e. in equation 3, see supplementary table S4), 499

malaria infected mice seem to have a lower threshold centered at 0.4 (Figure 2F). It is 500

speculated that the enlargement of malarial spleen, and the further spleen enlargement 501

after CQ treatment have modified the pore size in the splenic meshwork and altered 502

RBC microcirculation (2). Lowered spleen retention threshold could imply enlarged pore 503

size in malaria infected mice, which is consistent with several independent studies (58-504

60) using intravital microscopy and magnetic resonance imaging, which reported dilated 505

venous sinuses in the red pulp of enlarged spleen. 506

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

17

CONCLUSIONS 507

By employing a microfluidic single cell deformability cytometer, we investigated 508

the filterability or deformability of PVB and SMB from healthy and malaria infected mice. 509

Several interesting observations were made from the population-wide distribution of 510

single RBC deformability: 1) RBCs extracted from SMB are less deformable compared to 511

cells from PVB, suggesting passive splenic RBC retention of less deformable RBCs might 512

have occurred, though the retention rate is typically low for healthy test subjects; 2) CQ 513

has a general stiffening effect on hRBCs, uRBCs and iRBCs, which could result in an 514

increased splenic RBC retention of all RBCs. The retention is evident both from RBC 515

deformability measurement as well as spleen size quantification. It is likely that CQ 516

induced alteration in RBC microcirculatory behavior is attributed to increased oxidative 517

stress. 3) Increased splenic RBC retention strongly corresponds to anemic condition for 518

both healthy and malaria infected mice. Malarial anemia might be caused by the direct 519

impact of intense splenic RBC retention, rather than high parasite loading. These new 520

insights are expected to be useful in developing strategies to deal with malarial anemia, 521

which is one of the severe pathological outcomes of (chronic) malaria infection. 522

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

18

Acknowledgments 523

The authors would like to thank Dr. Subra Suresh for helpful discussions. Device 524

fabrications were carried out at MIT Microsystems Technology Laboratories. This work is 525

supported by the National Research Foundation (Singapore) through Singapore-MIT 526

Alliance for Research and Technology (SMART) Center (BioSyM and ID IRG). 527

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

19

References 528

529 1. Mebius RE, Kraal G. 2005. Structure and function of the spleen. Nat Rev Immunol 5:606-530

616. 531 2. del Portillo HA, Ferrer M, Brugat T, Martin-Jaular L, Langhorne J, Lacerda MVG. 2012. 532

The role of the spleen in malaria. Cellular Microbiology 14. 533 3. Schmidt EE, Macdonald IC, Groom AC. 1993. Comparative aspects of splenic 534

microcirculatory pathways in mammals: the region bordering the white pulp. Scanning 535 Microsc. 7:613-628. 536

4. Cesta MF. 2006. Normal Structure, Function, and Histology of the Spleen. Toxicologic 537 Pathology 34:455-465. 538

5. Schmidt EE, MacDonald IC, Groom AC. 1988. Microcirculatory pathways in normal 539 human spleen, demonstrated by scanning electron microscopy of corrosion casts. 540 American Journal of Anatomy 181:253-266. 541

6. Chen LT. 1978. Microcirculation of the spleen: an open or closed circulation?, vol. 201. 542 7. Buffet PA, Safeukui I, Milon G, Mercereau-Puijalon O, David PH. 2009. Retention of 543

erythrocytes in the spleen: a double-edged process in human malaria. Current Opinion 544 in Hematology 16:157-164. . 545

8. Suresh S, Spatz J, Mills JP, Micoulet A, Dao M, Lim CT, Beil M, Seufferlein T. 2005. 546 Connections between single-cell biomechanics and human disease states: 547 gastrointestinal cancer and malaria. Acta Biomater. 1:15-30. 548

9. Mills JP, Diez-Silva M, Quinn DJ, Dao M, Lang MJ, Tan KSW, Lim CT, Milon G, David PH, 549 Mercereau-Puijalon O, Bonnefoy S, Suresh S. 2007. Effect of plasmodial RESA protein 550 on deformability of human red blood cells harboring Plasmodium falciparum. 551 Proceedings of the National Academy of Sciences of the United States of America 552 104:9213–9217. 553

10. Bow H, Pivkin IV, Diez-Silva M, Goldfless SJ, Dao M, Niles JC, Suresh S, Han J. 2011. A 554 microfabricated deformability-based flow cytometer with application to malaria. Lab on 555 a Chip 11:1065-1073. 556

11. Diez-Silva M, Park Y, Huang S, Bow H, Mercereau-Puijalon O, Deplaine G, Lavazec C, 557 Perrot S, Bonnefoy S, Feld M, Han J, Dao M, Suresh S. 2012. Pf155/RESA protein 558 influences the dynamic microcirculatory behavior of ring-stage Plasmodium falciparum 559 infected red blood cells. Scientific Reports 2:614. 560

12. Park Y, Diez-Silva M, Popescu G, Lykotrafitis G, Choi W, Feld MS, Suresh S. 2008. 561 Refractive index maps and membrane dynamics of human red blood cells parasitized by 562 Plasmodium falciparum. Proceedings of the National Academy of Sciences 105:13730-563 13735. 564

13. Huang S, Undisz A, Diez-Silva M, Bow H, Dao M, Han J. 2013. Dynamic deformability of 565 Plasmodium falciparum-infected erythrocytes exposed to artesunate in vitro. Integrative 566 Biology. 567

14. Buffet PA, Milon G, Brousse V, Correas J-M, Dousset B, Couvelard A, Kianmanesh R, 568 Farges O, Sauvanet A, Paye F, Ungeheuer M-N, Ottone C, Khun H, Fiette L, Guigon G, 569 Huerre M, Mercereau-Puijalon O, David PH. 2006. Ex vivo perfusion of human spleens 570 maintains clearing and processing functions. Blood 107:3745-3752. 571

15. Safeukui I, Correas J-M, Brousse V, Hirt D, Deplaine G, Mulé S, Lesurtel M, Goasguen N, 572 Sauvanet A, Couvelard A, Kerneis S, Khun H, Vigan-Womas I, Ottone C, Molina TJ, 573 Tréluyer J-M, Mercereau-Puijalon O, Milon G, David PH, Buffet PA. 2008. Retention of 574

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

20

Plasmodium falciparum ring-infected erythrocytes in the slow, open microcirculation of 575 the human spleen. Blood 112:2520-2528. 576

16. Deplaine G, Safeukui I, Jeddi F, Lacoste F, Brousse V, Perrot S, Biligui S, Guillotte M, 577 Guitton C, Dokmak S, Aussilhou B, Sauvanet A, Cazals Hatem D, Paye F, Thellier M, 578 Mazier D, Milon G, Mohandas N, Mercereau-Puijalon O, David PH, Buffet PA. 2011. 579 The sensing of poorly deformable red blood cells by the human spleen can be mimicked 580 in vitro. Blood 117:88-95. 581

17. Buffet PA, Safeukui I, Deplaine G, Brousse V, Prendki V, Thellier M, Turner GD, 582 Mercereau-Puijalon O. 2011. The pathogenesis of Plasmodium falciparum malaria in 583 humans: insights from splenic physiology. Blood 117:381-392. 584

18. Looareesuwan S, Ho M, Wattanagoon Y, White NJ, Warrell DA, Bunnag D, Harinasuta T, 585 Wyler DJ. 1987. Dynamic Alteration in Splenic Function during Acute falciparum Malaria. 586 New England Journal of Medicine 317:675-679. 587

19. Chotivanich K, Udomsangpetch R, McGready R, Proux S, Newton P, Pukrittayakamee S, 588 Looareesuwan S, White NJ. 2002. Central Role of the Spleen in Malaria Parasite 589 Clearance. Journal of Infectious Diseases 185:1538-1541. 590

20. Lamikanra AA, Brown D, Potocnik A, Casals-Pascual C, Langhorne J, Roberts DJ. 2007. 591 Malarial anemia: of mice and men. Blood 110:18-28. 592

21. Dondorp AM, Angus BJ, Chotivanich K, Silamut K, Ruangveerayuth R, Hardeman MR, 593 Kager PA, Vreeken J, White NJ. 1999. Red blood cell deformability as a predictor of 594 anemia in severe falciparum malaria. The American Journal of Tropical Medicine and 595 Hygiene 60:733-737. 596

22. Price RN, Simpson JA, Nosten F, Luxemburger C, Hkirjaroen L, ter Kuile F, 597 Chongsuphajaisiddhi T, White NJ. 2001. Factors contributing to anemia after 598 uncomplicated falciparum malaria. The American Journal of Tropical Medicine and 599 Hygiene 65:614-622. 600

23. Sullivan DJ, Gluzman IY, Russell DG, Goldberg DE. 1996. On the molecular mechanism 601 of chloroquine's antimalarial action. Proceedings of the National Academy of Sciences 602 93:11865-11870. 603

24. Esposito A, Tiffert T, Mauritz JMA, Schlachter S, Bannister LH, Kaminski CF, Lew VL. 604 2008. FRET Imaging of Hemoglobin Concentration in Plasmodium falciparum-Infected 605 Red Cells. PLoS ONE 3:3780. 606

25. Fitch CD, Chevli R, Kanjananggulpan P, Dutta P, Chevli K, Chou AC. 1983. Intracellular 607 ferriprotoporphyrin IX is a lytic agent. Blood 62:1165-1168. 608

26. Carlton JM, Angiuoli SV, Suh BB, Kooij TW, Pertea M, Silva JC, Ermolaeva MD, Allen JE, 609 Selengut JD, Koo HL, Peterson JD, Pop M, Kosack DS, Shumway MF, Bidwell SL, 610 Shallom SJ, van Aken SE, Riedmuller SB, Feldblyum TV, Cho JK, Quackenbush J, 611 Sedegah M, Shoaibi A, Cummings LM, Florens L, Yates JR, Raine JD, Sinden RE, Harris 612 MA, Cunningham DA, Preiser PR, Bergman LW, Vaidya AB, van Lin LH, Janse CJ, Waters 613 AP, Smith HO, White OR, Salzberg SL, Venter JC, Fraser CM, Hoffman SL, Gardner MJ, 614 Carucci DJ. 2002. Genome sequence and comparative analysis of the model rodent 615 malaria parasite Plasmodium yoelii yoelii. Nature 419:512-519. 616

27. Negreiros RMdA, Makimoto FH, Santana LLO, Ferreira LCdL, Nakajima GS, Santos MCd. 617 2009. Experimental splenectomies and malaria in mice. Acta Cirurgica Brasileira 24:437-618 441. 619

28. Bapat D, Huang X, Gunalan K, Preiser PR. 2011. Changes in parasite virulence induced 620 by the disruption of a single member of the 235 kDa rhoptry protein multigene family of 621 Plasmodium yoelii. PLoS ONE 6:20170. 622

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

21

29. Stutte HJ. 1968. Nature of human spleen red pulp cells with special reference to sinus 623 lining cells. Cell and Tissue Research 91:300-314. 624

30. Hataba Y, Kirino Y, Suzuki T. 1981. Scanning Electron Microscopic Study of the Red Pulp 625 of Mouse Spleen. Journal of Electron Microscopy 30:46-56. 626

31. Klausner MA, Hirsch LJ, Leblond PF, Chamberlain JK, Klemperer MR, Segel GB. 1975. 627 Contrasting splenic mechanisms in the blood clearance of red blood cells and colloidal 628 particles. Blood 46:965-976. 629

32. Lewis M, Pfeil J, Mueller A-K. 2011. Continuous oral chloroquine as a novel route for 630 Plasmodium prophylaxis and cure in experimental murine models. BMC Res Notes 4:1-4. 631

33. Hagihara N, Walbridge S, Olson AW, Oldfield EH, Youle RJ. 2000. Vascular Protection by 632 Chloroquine during Brain Tumor Therapy with Tf-CRM107. Cancer Research 60:230-234. 633

34. Young LE, Platzer RF, Ervin DM, Izzo MJ. 1951. Hereditary Spherocytosis. Blood 6:1099-634 1113. 635

35. Safeukui I, Buffet PA, Deplaine G, Perrot S, Brousse V, Ndour A, Nguyen M, Mercereau-636 Puijalon O, David PH, Milon G, Mohandas N. 2012. Quantitative assessment of sensing 637 and sequestration of spherocytic erythrocytes by the human spleen. Blood 120:424-430. 638

36. Murphy SC, Fernandez-Pol S, Chung PH, Prasanna Murthy SN, Milne SB, Salomao M, 639 Brown HA, Lomasney JW, Mohandas N, Haldar K. 2007. Cytoplasmic remodeling of 640 erythrocyte raft lipids during infection by the human malaria parasite Plasmodium 641 falciparum. Blood 110:2132-2139. 642

37. FUDENBERG H, BALDINI M, MAHONEY JP, DAMESHEK W. 1961. The Body 643 Hematocrit/Venous Hematocrit Ratio and the "Splenic Reservoir". Blood 17:71-82. 644

38. Macomber PB, Sprinz H, Tousimis AJ. 1967. Morphological Effects of Chloroquine on 645 Plasmodium berghei in Mice. Nature 214. 646

39. Desai KR, Dattani JJ, Rajput DK, Moid N, Yagnik BJ, Highland HN, George LB. 2010. 647 Effect of chronic administration of chloroquine on the gastrocnemius muscle, spleen 648 and brain of Swiss albino mice. Asian Journal of Traditional Medicines. 649

40. Smith LP, Hunter KW, Oldfield EC, Strickland GT. 1982. Murine malaria: blood clearance 650 and organ sequestration of Plasmodium yoelii-infected erythrocytes. Infection and 651 Immunity 38:162-167. 652

41. Totino P, Magalhaes A, Silva L, Banic D, Daniel-Ribeiro C, Ferreira-da-Cruz M. 2010. 653 Apoptosis of non-parasitized red blood cells in malaria: a putative mechanism involved 654 in the pathogenesis of anaemia. Malaria Journal 9:350. 655

42. Low P, Waugh S, Zinke K, Drenckhahn D. 1985. The role of hemoglobin denaturation 656 and band 3 clustering in red blood cell aging. Science 227:531-533. 657

43. Chien S. 1987. Red Cell Deformability and its Relevance to Blood Flow. Annual Review of 658 Physiology 49:177-192. 659

44. MacDonald IC, Schmidt EE, Groom AC. 1991. The high splenic hematocrit: A rheological 660 consequence of red cell flow through the reticular meshwork. Microvascular Research 661 42:60-76. 662

45. MacDonald IC, Ragan DM, Schmidt EE, Groom AC. 1987. Kinetics of red blood cell 663 passage through interendothelial slits into venous sinuses in rat spleen, analyzed by in 664 vivo microscopy. Microvascular Research 33:118-134. 665

46. Nuchsongsin F, Chotivanich K, Charunwatthana P, Fausta O-S, Taramelli D, Day NP, 666 White NJ, Dondorp AM. 2007. Effects of Malaria Heme Products on Red Blood Cell 667 Deformability. The American Journal of Tropical Medicine and Hygiene 77:617-622. 668

47. Glaser R. 1979. The shape of red blood cells as a function of membrane potential and 669 temperature. The Journal of Membrane Biology 51:217-228. 670

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

22

48. Safeukui I, Buffet PA, Perrot S, Sauvanet A, Aussilhou B. 2013. Surface Area Loss and 671 Increased Sphericity Account for the Splenic Entrapment of Subpopulations of 672 Plasmodium falciparum Ring-Infected Erythrocytes. PLoS ONE 8:60150. 673

49. Reinhart WH, Chien S. 1986. Red cell rheology in stomatocyte-echinocyte 674 transformation: roles of cell geometry and cell shape. Blood 67:1110-1118. 675

50. Iyawe HOT, Onigbinde AO. 2004. Effect of an Antimalarial and a Micronutrient 676 Supplementation on Respiration Induced Oxidative Stress. Pakistan Journal of Nutrition 677 3:318-321. 678

51. Bhatia A CP. 1985. Chloroquine induces oxidative lysis of Plasmodium berghei 679 parasitized red blood cells. Ann Soc Belg Med Trop 65. 680

52. Liu TZ, Lin TF, Hung IJ, Wei JS, Chiu D. 1994. Enhanced susceptibility of erythrocytes 681 deficient in glucose-6-phosphate dehydrogenase to alloxan/glutathione-induced 682 decrease in red cell deformability. Life Sciences 55:55-60. 683

53. Chiu D, Liu T. 1997. Free radical and oxidative damage in human blood cells. Journal of 684 Biomedical Science 4:256-259. 685

54. Fitch CD, Russell NV. 2006. Accelerated Denaturation of Hemoglobin and the 686 Antimalarial Action of Chloroquine. Antimicrobial Agents and Chemotherapy 50:2415-687 2419. 688

55. Horký J VJ, Znojil V. 1978;. Comparison of life span of erythrocytes in some inbred 689 strains of mouse using 14C-labelled glycine. Physiol Bohemoslov 27:209-217. 690

56. Waugh RE, Narla M, Jackson CW, Mueller TJ, Suzuki T, Dale GL. 1992. Rheologic 691 properties of senescent erythrocytes: loss of surface area and volume with red blood 692 cell age. Blood 79:1351-1358. 693

57. Mohandas N, Gallagher PG. 2008. Red cell membrane: past, present, and future. Blood 694 112:3939-3948. 695

58. Freitas CRL, Barbosa Jr AA, Fernandes A, Andrade ZA. 1999. Pathology of the spleen in 696 hepatosplenic schistosomiasis. Morphometric evaluation and extracellular matrix 697 changes. Mem Inst Oswaldo Cruz. 94:815-822. 698

59. Ferrer M, Martin-Jaular, L., Calvo, M., del Portillo, H. A. 2012. Intravital Microscopy of 699 the Spleen: Quantitative Analysis of Parasite Mobility and Blood Flow. J. Vis. Exp. 59. 700

60. Bae K, Jeon KN. 2006. CT findings of malarial spleen. British Journal of Radiology 79:145-701 147. 702

703

704

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

23

Figure Legends 705 706

Figure 1. Deformability of red blood cells (RBC) extracted from SMB was compared to that from 707 PVB, in both healthy (A&B) and infected (C&D) mice. In the healthy mice, the velocity of hRBCs 708 in the spleen was significantly lower than that in peripheral (A, p<0.0001). Data was replotted 709 into histogram and a velocity threshold was then drawn at 0.65 (B). Similarly, in the infected 710 mice, splenic iRBCs and uRBCs were significantly less deformable than corresponding peripheral 711 cells (p<0.001, C). Data was replotted into histogram and a velocity threshold of 0.40 was drawn 712 (D). Significant spleen enlargement was observed for malaria infected mice (E). All deformability 713 measurements were performed using a microfluidic device shown in Figure S1B. Deformability 714 data are pulled from 6-9 mice per condition. 715 716 Figure 2. The effects of malaria infection (A&B) or/and antimalarial drug treatment (C-F) on RBC 717 deformability were investigated. When we extract RBCs from healthy mice (therefore hRBCs) 718 and RBCs from infected mice (therefore uRBCs and iRBCs), the mean velocity of hRBCs is slightly 719 higher than uRBCs (p<0.001), and considerably much higher than iRBCs (p<0.001, A). Data was 720 replotted into histogram for better illustration. When infected mice treated with CQ, the iRBCs 721 were found to be significantly less deformable (p<0.001, C). Data was replotted into histogram 722 to illustrate the shift in population-wide RBC velocity (D). In the spleen of infected mice, uRBCs 723 were also found to be significantly less deformable (p<0.001, E). A velocity threshold of 0.40 was 724 drawn in F. Significant spleen enlargement was observed for malaria infected mice after CQ 725 treatment, as compared to infected mice treating with PBS buffer only (G). Deformability data 726 are pulled from 6-9 mice per condition. It is also noted that the data for hRBC, uRBC and iRBC 727 shown in A-D (control group in C&D) are repetitive to easier comparison. 728 729 Figure 3. CQ effect on healthy RBCs in vivo (A&B). Whereas only very subtle decrease in RBC 730 deformability was observed in vivo peripheral blood (A), a noticeable shift in RBC deformability 731 profile (from unimodal to bimodal) was demonstrated in vivo splenic minced blood (B). 732 Deformability data are pulled from 6-9 mice per condition 733 734 Figure 4. Hemoglobin assay was performed for the venous blood extracted from healthy and 735 malarial infected mice (A). Infected mice after CQ treatment exhibited significantly larger spleen 736 in terms of mass and length (B). (p<0.05) 737 738

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

A C

1.21.41.61.8

ocity

Healthy Mice

1.2

1.4

1.6p < 0.001

Mice Infected with P Yoelii oc

ity

p < 0.001

Figure 1

0 20.40.60.81.01.2

Nor

mal

ized

Vel

o

<0 0001 0 2

0.4

0.6

0.8

1.0

Nor

mal

ized

Vel

o

10

SMB: iRBC200

300

PVB: hRBC

DB

0.00.2

Peripheral

N

Spleen

p <0.0001

0.0

0.2p < 0.001p < 0.001

uRBCs

N

SpleenPeripheraliRBCs uRBCs iRBCs

15

SMB:uRBC

0

5

SMB: iRBC

100

150

SMB

0

100

Freq

uenc

y

20

30

requ

ency

PVB: iRBC

0

5

10

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.80

50

Normalized Velocity

E

40

60

PVB:uRBC

0

10

20

FrE

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.60

20

40

Normalized VelocityMalaria infected

mice spleenHealthy

mice spleen

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

A C E1 2

1.4

1.6 p <0.001p <0.001RBCs treated with CQ: Peripheral

y 1 2

1.4

1.6RBCs treated with CQ: Spleen

ty

p<0.0011.41.61.8

healthy RBC v.s. infected RBC

y

p <0.001

Figure 2

0.4

0.6

0.8

1.0

1.2

Nor

mal

ized

Vel

ocity

0.4

0.6

0.8

1.0

1.2

orm

aliz

ed V

eloc

it

0.40.60.81.01.21.4

orm

aliz

ed V

eloc

it

30

PVB iRBC

B D F0.0

0.2

N

uRBCs iRBCs uRBCs iRBCsControl Group CQ Treated Group

p <0.0010.0

0.2

No

Control uRBCs CQ treated uRBCs

10

15

SMB: uRBC+CQ

10

iRBC CQ

0.00.2 p <0.001

No

hRBCs uRBCs iRBCs

60 0

10

20 PVB: iRBCs

10

15

20

Freq

uenc

y

SMB: uRBC0

5

20

uRBC+CQ

0

5iRBC+CQ

0

20

40

60

Freq

uenc

y PVB: uRBCs

G Malarial-Drug

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.60

5

Normalized Velocity

102030

PVB: iRBCs

0

10

Freq

uenc

y

50100150200250

PVB: hRBCs

Drug

20406080

PVB: uRBCs

010

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.80

Normalized Velocity

Malarial-Control

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.60

Normalized Velocity

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Figure 3

1 21.41.61.8

p>0.1p<0.001

RBCs treated with CQ: Peripheral

ocity

p<0.001A. B.

1 21.41.61.8

oc

ityRBCs treated with CQ: Spleen

p>0.1p<0.001

p<0.001

0.20.40.60.81.01.2

Nor

mal

ized

Vel

o

0.20.40.60.81.01.2

Nor

mal

ized

Vel

o

0.00.2

Control 5 7 9in Peripheral (days after CQ injection)

0.00.2

Day 5 Control Day 9Day 7

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Figure 4

1 0 Spleen mass and size of infected miceA. B.30 Healthy Infected (% of parasitemia)

0.20.40.60.81.0 p < 0.05

mas

s (g

) Day 4

20

25 Day 1 Day 5 +CQ

Day 1 (1~10%) Day 4 + PBS (>50%) Day 4 +CQ(<1%)

/dl)

p<0.01

p<0.01p<0.01

0.0

1234

gth

(cm

) p < 0.05

5

10

15

Hb

(g

p<0.01

01

leng

PBS-treated CQ-treated0

5

on March 8, 2018 by guest

http://iai.asm.org/

Dow

nloaded from