aor_1296 210..223

Thoughts and Progress

Impact of Hemoglobin Concentration andAffinity for Oxygen on Tissue

Oxygenation: The Case ofHemoglobin-Based Oxygen Carriers

Michele Samaja and Laura TerraneoDipartimento di Medicina, Chirurgia e Odontoiatria,

Università di Milano—San Paolo, Milano, Italy

Abstract: In patients undergoing exchange-transfusionwith hemoglobin (Hb)-based oxygen (O2) carriers(HBOC), native Hb coexists with newly transfused Hb.The two Hb types share the same arterial and venousPO2, but their affinities for O2 vary. A simple spreadsheetmodel is described aiming at evaluating the contributionof each Hb type to the overall O2 transport characteristicsas a function of the batch Hb concentration and O2 affin-ity in the HBOC solution, of the fraction of exchange-transfused blood/HBOC, and of the arterial PO2. Thismodel helps to yield a quantitative estimate of howtissues with high or low O2 extraction respond to thechanges cited above. The results show that the higher theexchange-transfusion ratio, the O2 transport to tissuesbecomes progressively impaired. However, this effect ismore critical at low batch Hb concentration and high O2

affinity of the HBOC, especially for tissues/organs withhigh O2 extraction, whereas the arterial PO2 does notappear as critical. Key Words: Hemoglobin—Oxygendelivery—Hypoxia—Hemoglobin-based oxygen carriers.

Hemoglobin (Hb), a red blood cell (RBC) protein,reversibly binds to the oxygen (O2) in the lungs torelease it to the tissues. The complex interaction withCO2, 2,3-diphosphoglycerate (DPG), Cl- (for bovineHb), and protons determines the Hb affinity for O2

(1), which is conveniently expressed by its P50, orPO2 at which half of Hb is bound to O2. By influenc-ing the arterial and venous Hb–O2 saturations (SaO2

and SvO2, mole/mole) at any arterial and venous O2

partial pressure (PaO2 and PvO2), the Hb–O2 affinitycontributes to tune the arteriovenous O2 difference(D[O2]a-v):

Δ[O ] [Hb] (S O S O ) (P O P O )2 a-v a 2 v 2 a 2 v= × − + × −α 2

(1)

[Hb] and a represent the blood Hb concentrationand the O2 solubility coefficient (1.4 ¥ 10-6 M/mm Hg(2)), respectively. Although unlikely to accuratelyyield the real D[O2]a-v, this relationship helps inassessing how changes in a single variable reflect intochanges in D[O2]a-v. When the blood flow (Q) isknown or assumed, this relationship translates intothe Fick’s equation, where VO2 represents the O2

consumption:

VO Q [O ]2 2 a-v= × Δ (2)

Whereas in a systemic situation Q and VO2 are,respectively, the cardiac output and the body O2 con-sumption, with “venous” referring to the mixedvenous return, in a local situation within an organ ortissue, Q and VO2 represent the blood flow andorgan/tissue O2 consumption, with “venous” referringto the end-capillary compartment. In general, a mod-erate P50 increase tends to diminish SvO2 withoutaffecting considerably SaO2. This increases D[O2]a-v

favoring the O2 delivery when the other factors areconstant. For example, increasing blood P50 by2 mm Hg allows either 10% Q reduction (less cardiacload) or higher PvO2 (less tissue hypoxia) (3).

Hb-based O2 carriers (HBOC) are emerging aspotential substitutes for blood in emergencysituations. In the blood of patients transfused withHBOC, Hb from HBOC (HbHBOC) coexists withnative RBC Hb (HbRBC). Despite sharing the samePaO2, PvO2, and Q, the two Hbs differ in their relativeabundance ([HbHBOC] and [HbRBC]) and O2 affinity(P50HBOC and P50RBC), hence SaO2 and SvO2. The two-compartment nature of such a situation can bemanaged by the following equation:

Δ[O ] Hb ] (S O S O ) [Hb ]

(S O S Oa-v RBC a 2RBC v 2RBC HBOC

a 2HBOC v 2

2 = × − + ×−

[

HHBOC a 2 v 2) (P O P O )+ × −α(3)

With some considerable exception, the variousHBOCs frequently display a high Hb–O2 affinity,with P50 as low as 5 mm Hg.This feature is perceivednegatively because the high Hb–O2 affinity is pre-dicted to impair the O2 delivery: HbHBOC binds to the

doi:10.1111/j.1525-1594.2011.01296.x

Received July 2010; revised February 2011.Address correspondence and reprint requests to Professor

Michele Samaja, Dipartimento di Medicina, Chirurgia e Odontoia-tria, San Paolo, Università di Milano, Via di Rudinì 8, Milano20142, Italy. E-mail: michele.samaja@unimi.it

Artificial Organs36(2):210–223, Wiley Periodicals, Inc.© 2011, Copyright the AuthorsArtificial Organs © 2011, International Center for Artificial Organs and Transplantation and Wiley Periodicals, Inc.

210

O2 strongly in the lungs and releases it slowly to thetissues. In the lack of a suitable HBOC to address thisissue experimentally without overlapping phenom-ena such as vasoconstriction, extravasation, increasedsystemic vascular resistance, and methemoglobin for-mation, the computer simulation object of this reportuses the two-compartment modeling of Eq. (3) topredict the effects of changing some selected vari-ables on D[O2]a-v.

METHODS

To implement the model, we used the O2 equilib-rium curves (OECs) of HbRBC and propositus HbHBOC

with a PO2 resolution of 0.25 mm Hg in the range0–100 mm Hg. The OEC for HbRBC was obtainedassuming normal values for CO2, pH, and DPG(P50 = 28 mm Hg) (4).The OECs for a high and a lowO2 affinity HbHBOC (MP4 and aaHb, P50 = 5 and30 mm Hg, respectively) were provided fromSangart, Inc. (San Diego, CA, USA). The OECs forHBOCs with P50 = 20 and 40 mm Hg were obtainedfrom intrapolation/extrapolation of the mentionedones.

D[O2]a-v was calculated from Eq. (3) via an Excelspreadsheet freely available from the authors. Theinputs are the OECs for HbRBC and HbHBOC andtheir blood Hb concentrations [HbRBC] and[HbHBOC]. The total O2 content (chemically bound toHb plus physically dissolved) in the arterial blood([O2]a) at the selected PaO2 is calculated from theOEC as:

[O ] [Hb ] S O [Hb ] S OP O

2 a RBC a 2RBC HBOC a 2HBOC

a 2

= × + × +×α (4)

The calculation of the total O2 content in venousblood ([O2]v) is subject to experimental limitationsfor the known obstacles in measuring end-capillaryand mixed venous return PvO2. Therefore, amatrix of [O2]v values is generated as a function ofPvO2 in the range 0-PaO2 by the following equation,where SvO2 is obtained at each PvO2 using the OEC:

[O ] [Hb ] S O [Hb ] S OP O

2 v RBC v 2RBC HBOC v 2HBOC

v 2

= × + × +×α (5)

Finally, D[O2]a-v is calculated at any PvO2 and [O2]v:

Δ[O ] [O [O2 a-v 2 a 2 v= −] ] (6)

Therefore, the final matrix reports D[O2]a-v as afunction of PvO2 in the range 0-PaO2 mm Hg.

RESULTS

For model testing, consider the case of a 70-kgpatient undergoing isovolumetric exchange-transfusion (XT) with MP4 (P50 = 5 mm Hg) withbatch [Hb] = 4 g/dL.The XT ratio varies from 0% (noXT) to 100% (blood entirely replaced by HBOC).[HbHBOC] and [HbRBC] are calculated at each XT ratioassuming blood volume = 5 L. Figure 1 reports theOECs (panel A), the total blood O2 content (panelB), and the relationship between D[O2]a-v and PvO2 atvarying XT ratios (panel C). An interpretationfollows. When PvO2 increases, SvO2 approaches SaO2,and hence, D[O2]a-v tends to zero. By contrast, whenPvO2 approaches 0, D[O2]a-v depends on total [Hb], or[HbHBOC] + [HbRBC], because O2 is unloaded indepen-dently of the Hb–O2 affinity. At constant batch[HbHBOC] and blood volume, increasing XT ratiodecreases [HbHBOC] + [HbRBC], and hence D[O2]a-v. Atintermediate PvO2, D[O2]a-v depends on [Hb] and P50.Should all Hb occur as HbHBOC (XT = 100%) andP50HBOC is low, then a very low PvO2 is needed tounload O2. When Hb is present as both HbRBC andHbHBOC, an intermediate situation occurs.As a whole,when batch [Hb] and P50HBOC are low, a progressivelylower PvO2 is needed to extract O2 with increasingXT ratio.

Figure 2 predicts D[O2]a-v, assuming the above caseand after varying systematically one of the followingvariables: batch [Hb] in HBOC, which reflects intovarying [HbHBOC] (panel A), P50HBOC (panel B) andPaO2 (panel C), while keeping constant the XT ratio(50%, hence [HbRBC]) and P50RBC.As all the variablesaffect D[O2]a-v, the knowledge of PvO2 represents themost crucial point in this modeling.

As it is intuitive that PvO2 depends primarily ontissue O2 extraction, a likely value for PvO2 wasderived from the reported tissue/organ VO2, Q, andD[O2]a-v (5), assuming a resting male subject withVO2 = 0.25 L/min, PaO2 = 95 mm Hg, P50RBC = 28mm Hg, and blood [Hb] = 16 g/dL. The estimatedPvO2 was then used to interpolate D[O2]a-v in Fig. 2.Figure 3 reports the assumed PvO2 values and theimpact of [HbHBOC] PaO2 and P50HBOC in the consid-ered organs while maintaining constant the XT ratio(50%), [Hb]RBC, and P50RBC. It appears that organswith high O2 extraction respond to the presence ofHbHBOC in the blood by keeping lower PvO2 organswith low O2 extraction. Thus, in patients undergoingXT with HBOC, high O2 extraction organs suffer ahigher degree of hypoxia than low O2 extractionorgans. Furthermore, if the slope of the curvesindexes the interdependence of the variablesrepresented on the axes, it appears that the

THOUGHTS AND PROGRESS 211

Artif Organs, Vol. 36, No. 2, 2012

positive effect of increasing [Hb]HBOC is evident onlywhen P50HBOC is high. By contrast, PaO2 does notappear as crucial because its value lies above thePO2 range where O2 unloading by Hb might haveany influence.

DISCUSSION

The model is limited by the following assumptions:

1 Absence of changes in the ventilation rate, or con-stant PaO2. However, the effect of varying PaO2 onD[O2]a-v does not appear as relevant as those due tochanges in other variables.

O2 equilibrium curve

O2 equilibrium curve

0 25 50 75 1000.0

0.2

0.4

0.6

0.8

1.0

28

5HBOC

Blood

P50, mm Hg

PO2, mm Hg

PO2, mm Hg

Hb-

O2

satu

rati

on

0 25 50 75 1000.0

2.5

5.0

7.5

10.0 0

25

50

75

100

XT, %

Tota

l [O

2], m

M

O2 delivery

0 25 50 75 1000

2

4

6

8

10 0

25

50

75

100

XT=0-100%P50HBOC=5 mm HgP50RBC=28 mm Hg[Hb]HBOC=0-8 g/dl[Hb]RBC=16-0 g/dlPaO2=95 mm Hg

XT, %

PvO2, mm Hg

[O2]

a-v,

mM

A

B

C

FIG. 1. O2 delivery in a 70-kg patient with blood volume = 5 L,undergoing exchange-transfusion (XT) with high O2-affinityHBOC. (A) O2 equilibrium curves (SO2 vs. PO2) of HbRBC

(P50 = 28 mm Hg) and HbHBOC (P50 = 5 mm Hg). (B) Total O2

content vs. PO2 assuming isovolumetric XT (batch [Hb] = 4 g/dL)for 0% (no XT), 25%, 50%, 75%, and 100% (all blood replacedwith HBOC). [HbRBC] and [HbHBOC] are thus varied stepwise16-to-0 and 0-to-8 g/dL, respectively. (C) O2 delivery, or [O2]a-[O2]v, assuming PaO2 = 95 mm Hg.

O2 delivery

O2 delivery

O2 delivery

0 25 50 75 1000

2

4

6

8

10

0

2

4

6

8XT=50%P50HBOC=5 mm HgP50RBC=28 mm Hg[Hb]HBOC=0-6 g/dl[Hb]RBC=8 g/dlPaO2=95 mm Hg

[HbHBOC], g/dl

PvO2, mm Hg

PvO2, mm Hg

PvO2, mm Hg

[O2]

a-v,

mM

[O2]

a-v,

mM

[O2]

a-v,

mM

0 25 50 75 1000

2

4

6

8

10

0

2

4

6

8

10

510202840XT=50%

P50HBOC=5-40 mm HgP50RBC=28 mm Hg[Hb]HBOC=2 g/dl[Hb]RBC=8 g/dlPaO2=95 mm Hg

P50HBOC, mm Hg

0 25 50 75 100

95

85

7565

55

XT=50%P50HBOC=5 mm HgP50RBC=28 mm Hg[Hb]HBOC=2 g/dl[Hb]RBC=8 g/dlPaO2=55-95 mm Hg

PaO2, mm Hg

A

B

C

FIG. 2. O2 delivery at varying batch [Hb] and P50 in the HBOCand oxygenation state. (A) Effect of varying [HbHBOC] in the range0→8 g/dL in the circulation, which is equivalent to 0→12 g/dL inthe batch, at constant P50RBC, P50HBOC, [HbRBC], and PaO2 (seeinset). (B) Effect of varying P50HBOC in the range 5–40 mm Hg atconstant P50RBC, [HbHBOC], [HbRBC], and PaO2 (see inset). (C)Effect of varying PaO2 in the range 55–95 mm Hg at constantP50RBC, P50HBOC, [HbHBOC], and [HbRBC] (see inset).

THOUGHTS AND PROGRESS212

Artif Organs, Vol. 36, No. 2, 2012

2 Constant level of the Hb allosteric factors pH,PCO2, 2,3-DPG, and Cl- (for bovine Hb) through-out the observation, which translates into constantP50RBC and P50HBOC. In human blood under stan-

dard conditions, P50 is relatively constant but sub-jected to fluctuations arising in response to changesin the mentioned allosteric effectors as well asother factors that determine the actual P50 in vivo,

P50HBOC=5 mm Hg P50HBOC=40 mm Hg

P50HBOC=5 mm Hg P50HBOC=40 mm Hg

0 2 4 6 80

10

20

30

40

50

60

70

80

[HbHBOC], g/dL [HbHBOC], g/dL

PvO

2, m

m H

gP

vO2,

mm

Hg

PvO

2, m

m H

g

0 2 4 6 8

Brain

Heart

Kidneys

Legs

Total

50 60 70 80 900

10

20

30

40

50

60

70

80

PaO2, mm Hg PaO2, mm Hg

50 60 70 80 90

Brain

Heart

Kidneys

Legs

Total

[Hb]HBOC=1 g/dL [Hb]HBOC=4 g/dL

0 10 20 30 400

10

20

30

40

50

60

70

80

P50HBOC, mm Hg P50HBOC, mm Hg

0 10 20 30 40

Brain

Heart

Kidneys

Legs

Total

A

B

C

FIG. 3. O2 delivery in various organs. (A)Effect of [HbHBOC] on the PvO2 needed todeliver the amount of O2 indicated in thetext at constant values of [HbHBOC] (4 g/dL), [HbRBC] (8 g/dL), P50RBC (28 mm Hg),and PaO2 (95 mm Hg). P50HBOC is either 5(left) or 40 mm Hg (right). (B) Effect ofPaO2 on the PvO2 needed to deliver theamount of O2 indicated in the text at con-stant values of [HbHBOC] (4 g/dL), [HbRBC](8 g/dL) and P50RBC (28 mm Hg). P50HBOC

is either 5 (left) or 40 mm Hg (right). (C)Effect of P50HBOC on the PvO2 neededto deliver the amount of O2 indicatedin the text at constant values of PaO2

(95 mm Hg), P50RBC (28 mm Hg), and[HbRBC] (8 g/dL). [HbHBOC] is either 1 (left)or 5 g/dL (right). The assumed valuesfor PvO2 for brain, heart, kidneys, legs,and total body are 44, 34, 70, 41, and47 mm Hg, respectively.

THOUGHTS AND PROGRESS 213

Artif Organs, Vol. 36, No. 2, 2012

including temperature, hydrostatic pressure, intra-capillary Bohr effect, and varying RBC distribu-tion (e.g., “old” and “young” RBCs coexisting inthe same blood stream) (6). Some of these changesmight, in principle, affect P50HBOC as well.

3 No loss of HBOC or RBC from the circulation,or constant [HbHBOC] and [HbRBC], whereas Hbextravasation is known to be a critical phenom-enon in HBOC experimentation.

4 Negligible effects of nitric oxide (NO), anendothelium-derived relaxing factor with anumber of effects in somatic cells. NO stimulatessoluble guanylate cyclase to form cyclic guanosinemonophosphate (cGMP), a messenger that medi-ates a large number of effects in the microcircula-tion in vivo. Current investigations also address thehypothesis that Hb is a NO transporter (7) thatresponds to changes in the Hb oxygenation state(8) and helps NO export out of the RBC promot-ing local vasorelaxation and improving VO2 uponHb deoxygenation (9).

5 Constant RBC capillary transit time with negligibleblood flow autoregulation. However, recent workaddressed cerebral blood flow autoregulation as asensitive mechanism to tune the blood O2 transportin response to the body/organ needs (10,11).

6 The described model cannot assess where exactlyO2 unloading in the capillary bed occurs, e.g., in theprecapillary area or elsewhere, while it is wellappreciated that O2 unloaded in the precapillarycontributes to vasoconstriction rather than tissueoxygenation (12).

It is worthwhile to note that the unavailability of invivo data obtained with HBOC in the place of humanblood led to plot Figs. 1C and 2 in a PvO2 range thatexceeds the range of clinically or physiologically rel-evant values.

Inadequate O2 delivery invariably leads to hypoxia,which might be lethal if severe and/or sustained.But inthe case of sublethal hypoxia, cells are enabled toresist by overexpression of the hypoxia-induciblefactor 1a (HIF-1a),a ubiquitarian molecule that regu-lates the transcription of a variety of downstreamgenes coding for proteins that elicit hypoxia adapta-tion (13). Besides being a hypoxia marker, HIF-1a isthus pivotal to orchestrate the cell response tohypoxia.Although this feature is not shared equally bythe body organs (14), HIF-1a accumulates in hypoxictissues in vivo,and such accumulation might be criticalto elicit a useful response to hypoxia, as, for example,triggering angiogenesis, resisting to apoptosis, increas-ing the antioxidant defense, and others. It is thus amatter for further investigation whether tissue

hypoxia must be avoided or may instead represent away to induce resistance to stress. Most likely, lack ofO2 is not the only trigger for HIF-1a overexpression,and not all the responses triggered by hypoxia aremediated by HIF-1a. Nevertheless, the HIF-1apathway is prominent in determining the cell/tissueability to respond to hypoxia, and its overexpressionmay be crucial to allow greater protection in HBOC-perfused tissues. It was observed that rats XT with apolymerized bovine Hb with P50 of 46 mm Hg andstock Hb concentration of 13 g/dL display attenuatedHIF-1a response in the kidney (15), despite increasedHIF-1a in hypoxic bovine aortic endothelial cellsincubated in the presence of diaspirin cross-linked Hb(16).That outcome was predicted by the present mod-eling (right plot in Fig. 3C), as kidneys perfused undercomparable conditions are far from being hypoxic,thereby providing an explanation for the low HIF-1aresponse in that organ.

CONCLUSION

The mathematical model enables predicting howbatch [Hb] and P50 in HBOC, as well as the PaO2,cooperate in tuning the O2 delivery to tissues inpatients undergoing XT with HBOC.The results showthat the paradigm “the higher the HBOC-O2 affinity,the less efficient O2 delivery becomes” might some-times be an oversimplification. If the slopes of thecurves in Fig. 3 provide an indication of the depen-dency of the O2 delivery on either [Hb]HBOC or PaO2, itappears that batch [Hb] becomes a critical factor onlywhen P50HBOC lies in the upper range. The PaO2

appears less critical than batch [Hb] and P50. Further-more, the O2 extraction by the single organ representsan additional factor to be considered because organswith high O2 extraction as the heart and brain arepredicted to respond to [Hb]HBOC and P50HBOC to aminor degree than organs with low O2 extraction asthe kidneys because of fluctuations in the local PvO2.In conclusion, it is not only the P50 that matters, butrather, the complex interaction among P50, [Hb], andPvO2 in the various organs that need to be taken intoaccount when assessing the O2 delivery efficiency ofHBOC.

Acknowledgments: We thank Drs. Kim Vandegriffand Ashok Malavalli (Sangart, Inc.) for helpful dis-cussion and providing O2 affinity data related to MP4and aaHb.

REFERENCES

1. Rovida E, Niggeler M, Carlone S, Samaja M. Carboxyhemo-globin and oxygen affinity of human blood. Clin Chem1984;30:1250–1.

THOUGHTS AND PROGRESS214

Artif Organs, Vol. 36, No. 2, 2012

2. Roughton F, Severinghaus J. Accurate determination of O2dissociation curve of human above 98.7% saturation with dataon O2 solubility in unmodified human blood from 0° to 37°C.J Appl Physiol 1973;35:861–9.

3. Samaja M, di Prampero P, Cerretelli P. The role of 2,3-DPG inthe oxygen transport at altitude. Respir Physiol 1986;64:191–202.

4. Winslow R, Samaja M, Winslow N, Rossi Bernardi L, ShragerR. Simulation of continuous blood O2 equilibrium curve overthe physiologic pH, DPG and pCO2 range. J Appl Physiol1983;54:524–9.

5. Folkow B, Neil E. Part II. The Heart and Circulation. In:Keele CA, Neil E, Joels N, eds. Samson Wright’s AppliedPhysiology. New York, Toronto: Oxford University Press, Inc,1982;65–154.

6. Samaja M, Crespi T, Guazzi M, Vandegriff K. Oxygen trans-port in blood at high altitude: role of the hemoglobin-oxygenaffinity and impact of the phenomena related to hemoglobinallosterism and red cell function. Eur J Appl Physiol2003;90:351–9.

7. Jia L, Bonaventura C, Bonaventura J, Stamler J.S-nitrosohemoglobin: a dynamic activity of blood involved invascular control. Nature 1996;380:221–6.

8. McMahon T, Stone A, Bonaventura J, Singel D, Stamler J.Functional coupling of oxygen binding and vasoactivityin S-nitrosohemoglobin. J Biol Chem 2000;275:16738–45.

9. McMahon T, Moon R, Luschinger B, et al. Nitric oxide in thehuman respiratory cycle. Nat Med 2002;8:711–7.

10. Ulatowski JA, Bucci E, Razynska A, Traystman RJ, KoehlerRC. Cerebral blood flow during hypoxic hypoxia with plasma-based hemoglobin at reduced hematocrit. Am J Physiol1998;274:H1933–42.

11. Koehler RC, Fronticelli C, Bucci E. Insensitivity of cerebraloxygen transport to oxygen affinity of hemoglobin-based oxygen carriers. Biochim Biophys Acta 2008;1784:1387–94.

12. Intaglietta M, Johnson P, Winslow R. Microvascular and tissueoxygen distribution. Cardiovasc Res 1996;32:632–43.

13. Manalo DJ, Rowan A, Lavoie T, et al. Transcriptional regula-tion of vascular endothelial cell responses to hypoxia byHIF-1. Blood 2005;105:659–69.

14. Bianciardi P, Fantacci M, Caretti A, et al. Chronic in vivohypoxia in various organs: hypoxia-inducible factor-1alphaand apoptosis. Biochem Biophys Res Commun 2006;342:875–80.

15. Manalo DJ, Buehler PW, Baek JH, Butt O, D’Agnillo F,Alayash AI. Acellular haemoglobin attenuates hypoxia-inducible factor-1alpha (HIF-1alpha) and its target genes inhaemodiluted rats. Biochem J 2008;414:461–9.

16. Yeh LH, Alayash AI. Effects of cell-free hemoglobin onhypoxia-inducible factor (HIF-1alpha) and heme oxygenase(HO-1) expressions in endothelial cells subjected to hypoxia.Antioxid Redox Signal 2004;6:944–53.

aor_1307 215..228

THOUGHTS AND PROGRESS 215

Artif Organs, Vol. 36, No. 2, 2012