Original Article
Original Article
Prehemolytic impact of phenothiazine drugs on the attachment of spectrin network in red blood cells
expand article infoBoyana Parvanova, Bilyana Tacheva, Ivan Ivanov
‡ Trakia University, Stara Zagora, Bulgaria
Open Access


Introduction: Chlorpromazine, thioridazine, and trifluoperazine are phenothiazine drugs that cause colloid-osmotic hemolysis of human erythrocytes by unknown mechanism. To clarify this mechanism, the impact of these drugs on the βsp (1.4 MHz) and γ1sp (9 MHz) dielectric relaxations was investigated. Each relaxation was shown to reduce its strength on the severing specific bridge that connects the spectrin network with the lipid membrane. For βsp relaxation, this is the spectrin-actin-glycophorin C bridge while for γ1sp relaxation this is the spectrin-ankyrin-band 3 bridge.

Aim: To elucidate the mechanisms of the effects of phenothiazine drugs in prehemolytic concentrations on the red blood cell plasma membrane using scanning temperature-dependent (thermal) differential dielectric spectroscopy.

Materials and methods: Erythrocytes were isolated from freshly drawn blood and 100 μl of them were suspended in 1 ml isotonic solution of 10 mM NaCl and mannitol (working medium) containing the indicated concentration of the drug for 10 min at 23°C. The treated erythrocytes were isolated, suspended in working medium, hematocrit 0.55, and heated (heating rate 1.5°C/min) above the denaturation temperature of spectrin (TA≈49.5°C) in order to obtain the differential dielectric spectroscopy data. The complex admittance, Y* = Y’+j.Y”, of the tested suspensions was continuously measured and separated into its real (Y’) and imaginary (Y”) parts using Solartron 1260 Impedance Frequency Analyzer.

Results: At pre-hemolytic concentrations, each drug inhibited these two relaxations, predominantly the γ1sp relaxation. The results could be interpreted in terms of a sigmoid effect of the drugs on the spectrin-ankyrin-band 3 bridge severing it at concentration just prior to the start of massive hemolysis.

Conclusions: The study points at the possible mechanism of erythrocyte damage after treatment with phenothiazine drugs at prehemolytic concentrations. This is probably due to the disruption of the bridges between the phospholipid bilayer and the submembrane spectrin network.


dielectric relaxations, drug-induced hemolysis, erythrocyte membrane


The plasma membrane of human erythrocytes consists of a network of mainly spectrin tetramers that supports the lipid membrane and contains two major integral proteins (band 3 and glycophorin C).[1] The spectrin network (membrane skeleton, MS) is attached to the lipid membrane by the so-called spectrin-ankyrin-band 3 bridge and by the spectrin-actin-glycophorin C bridge.

Chlorpromazine (CPZ), thioridazine (TRZ) and trifluoperazine (TFP) are tricyclic phenothiazine drugs used in the treatment of mental anxiety and disorders.[2] These drugs exhibit a two-phase action on erythrocyte membranes. At low concentrations, they stabilize the membranes and cause stomatocytosis of erythrocytes[3, 4], while at high concentrations, they induce colloid-osmotic hemolysis[5] by an unknown mechanism. To study this mechanism, we investigated the impact of CPZ, TRZ, and TFP on recently detected βsp and γ1sp dielectric relaxations on the spectrin network of erythrocytes.[6, 7] A recent report[6] has shown that βsp relaxation is inhibited by specific severing of spectrin-actin-glycophorin C bridge while γ1sp relaxation was inhibited by specific severing of spectrin-ankyrin-band 3 bridge. Based on this, the presented results could be interpreted by a sigmoid effect of studied drugs on the spectrin-ankyrin-band 3 bridge severing it just prior to the onset of hemolysis.


The aim of this study was to elucidate the mechanisms of the effects of phenothiazine drugs in pre-hemolytic concentrations on the red blood cell plasma membrane using thermal dielectric spectroscopy.

Materials and methods


NaCl, mannitol, chlorpromazine, thioridazine, and trifluoperazine were purchased from Sigma Chemicals Co, St. Louis, MO, USA.

Treatment of erythrocytes with phenothiazine drugs

Erythrocytes were isolated from freshly drawn blood as described previously.[6, 7] 100 μl of the erythrocytes were suspended in 1 ml isotonic solution of 10 mM NaCl and mannitol (working medium) containing the indicated concentration of the drug for 10 minutes at room temperature.‌[3] After the treatment, the level of hemolysis was evaluated spectrophotometrically at 600 nm. Prior to heating, the cells were isolated and re-suspended in the working medium, hematocrit 0.55.

Scanning temperature-dependent differential dielectric spectroscopy of erythrocyte suspensions at the spectrin denaturation temperature, TA

The obtained suspensions were immediately heated (from 30°C to 60°C with a heating rate of 1.5°C/min) across the TA (49.5°C) in order to obtain the differential dielectric spectroscopy at 16 frequencies between 20 kHz and 15 MHz with an integration time of 0.5 s data associated with the denaturation of spectrin as previously described.‌[6, 7] The complex admittance, Y* = Y’+j.Y”, of tested suspension was continuously measured and separated into its real (Y’) and imaginary (Y”) parts using Solartron 1260 Impedance Frequency Analyzer.

The complex admittance, Y*, of heated suspension has been shown to change sharply at TA and the obtained changes strongly depended on frequency (f).[6, 7] These changes ΔY’(f) and ΔY”(f), were corrected for the linear effect of temperature and were defined as the value at the native minus the value at the denatured state of spectrin. They were ascribed to the spectrin’s admittance contribution that was eliminated at TA. The complex plain plot, ΔY” vs. ΔY’, depicted two perfect semicircle arcs (Fig. 1A), one placed above and another one below the real axis.[6, 7] The upper arc expressed the βsp relaxation associated with a piezo effect powered by the electrostriction of the lipid membrane. The lower arc revealed the γ1sp relaxation associated with the resonance of electric field with the natural oscillations of dipoles (segments) of spectrin network.

Figure 1.

Complex plain plot (▲) of the admittance contribution of spectrin network of suspended erythrocytes (A). Curved arrow indicates the increase in frequency from 20 kHz to 15 MHz. Arrows indicate the characteristic frequencies of βsp and γ1sp dielectric relaxations. Open triangles (Δ) indicate the model plot, Y” vs. Y’, of the complex admittance, Y*, of equivalent electric circuit (B).

Model representation of βsp and γ1sp relaxations in erythrocytes

To obtain quantitative description of relaxations, we used an adequate electric model (Fig. 1B). The Rβsp and Cβsp of upper circuit represented the best-fit values for the βsp relaxation, while the Rγ1sp and Cγ1sp of lower circuit were the best-fit values for the γ1sp relaxation. The complex admittance, Y* = Y’+jY”, of the model circuit was obtained by iteration and the model plot, Y” vs. Y’, depicted two semicircles which pertained to the admittance change plot, ΔY” vs. ΔY’, of spectrin network as previously reported.[7] The Rβsp, Cβsp, Rγ1sp, and Cγ1sp parameters linearly depended on hematocrit when its values varied between 0.10 and 0.60. Hence, it was convenient to use the ratios of these parameters. Such are the energy dissipation ratio, −Rγ1sp/Rβsp and the energy storage ratio, −Cγ1sp/Cβsp, which represent the amount of energy dissipated and stored, respectively, on spectrin during the γ1sp relaxation relative to that in βsp relaxation. Compared to control erythrocytes, the −Rγ1sp/Rβsp ratio increased, while the −Cγ1sp/Cβsp ratio decreased strongly in case the spectrin-band 3 attachment bridge was severed.[6, 7]


Fig. 2 shows the effects produced by various concentrations of TFP on the complex plot, ΔY” vs. ΔY’, of spectrin’s admittance contribution for erythrocytes pre-treated with TFP. The TFP treatment induced either non- or insignificant hemolysis. Nevertheless, the ΔY” vs. ΔY’ plot demonstrated that TFP inhibited both relaxations whereat the γ1sp relaxation was much stronger reduced compared to βsp relaxation. For each concentration of TFP, the model plot was also obtained (not shown) and the best-fit values of its parameters are given in Table 1.

The data in Table 1 indicate that at concentrations up to 200 μM TFP weakly and linearly changed the model parameters, Rβsp and Cβsp, for the βsp relaxation. By contrast, for the γ1sp relaxation these parameters demonstrated sigmoid variation being almost constant up to 150 μM and sharply changing at 200 μM (Fig. 2, right). These results suggest that TFP-treatment of erythrocytes induced predominant detachment of spectrin-ankyrin-band 3 bridge. At the same time, the TFP-induced hemolysis was zero at concentrations below 150 μM, weak at 200 μM, and complete at 250 μM (data not shown).

The erythrocytes treated with TRZ and CPZ produced similar results to those reported above for TFP (not shown). Compared to TFP, the latter drugs differed by the concentrations at which they induced hemolysis and specifically inhibited the γ1sp relaxation. Up to the concentrations, where hemolysis became noticeable, the parameters of βsp relaxation changed weakly and linearly while those of γ1sp relaxation sustained marked sigmoid changes suggesting disruption of band 3-ankyrin-spectrin attachment bridge followed by massive hemolysis. The prehemolytic concentration interval was 200-250 μM for TFP, 400-450 μM for TRZ, and 1000-1200 μM for CPZ in line with the previous report.[3]

Figure 2.

Left: complex plain plot of spectrin’s admittance contribution for erythrocytes treated by TFP at concentrations 0 (▲), 150 μM (□), and 200 μM (○). Right: Effect of TFP concentration on the −Rγ1sp/Rβsp and −Cγ1sp/Cβsp ratios of TFP-treated erythrocytes.

Table 1.

Model study of the effect produced by TFP on the βsp and γ1sp relaxations in erythrocytes. Only the mean values are shown, the deviations were less than 10% of the mean values. Both relaxations are expressed by their characteristic frequencies, fβsp and fγ1sp, and the best-fit values of their RC circuits (Fig. 2B)

Concentration of TFP (μM) R βsp (kOhm) C βsp (pF) fβsp (MHz) R γ1sp (kOhm) C γ1sp (pF) f γ1sp (MHz) Rγ1sp/Rβsp Cγ1sp/Cβsp
0 −5.9 −23 1.2 6.7 2.5 8.0 1.13 0.11
75 −7.1 −19.4 1.2 9.2 2.2 7.5 1.3 0.11
150 −10.1 −14 1.1 15.4 1.7 6.5 1.52 0.12
200 −10.5 −13 1.2 27.1 0.74 8.0 2.57 0.06


The causes for destabilization of the erythrocyte membrane by phenothiazine drugs could be multiple. However, one of them could be elucidated taking into account the accompanying changes in the −Rγ1sp/Rβsp and −Cγ1sp/Cβsp ratios. According to a previous study[5], these changes suggest that spectrin-ankyrin-band 3 bridge was specifically severed just prior to the onset of massive hemolysis. This outcome could be attributed to the predominant incorporation of positively charged phenothiazines into the inner, negatively charged, leaflet of the membrane bilayer[3, 4], thereby disturbing the balance of spectrin-lipid membrane interactions. The latter conclusion is supported by the report[8] that the dissociation of spectrin-ankyrin-band 3 bridge leads to the decomposition of spectrin tetramers into constituent dimers impairing the mechanical function of spectrin network.

There are another two findings in line with the above conclusion. Electron microscopy of CPZ-treated erythrocytes has indicated almost two times greater width of their plasma membranes compared to intact erythrocytes.[5] This could be due to the increased number of spectrin dimers as each dimer is about 100 nm long and could be connected to the membrane only at one of its two ends. Freeze-fracture electron microscopy of the membrane of CPZ-treated erythrocytes revealed multiple groups of aggregated intramembrane particles (integral proteins) into the surrounding particle-free patches of bilayer membrane.[‌5] In general, such intramembrane particle aggregation is regarded as a consequence of the structural disturbance of MS.


This study suggests a possible mechanism responsible for the disturbance of MS and the membrane of erythrocytes by their treatment with prehemolytic concentrations of phenothiazine drugs.


This study was supported by the Medical Faculty of Trakia University of Stara Zagora, Bulgaria, project grant No. 3/2022.


  • 1. De Oliveira S, Saldanha C. An overview about erythrocyte membrane. Clin Hemorheol Microcirc 2010; 44(1):63–74.
  • 2. Lieberman JA, Phillips M, Gu H, et al. Atypical and conventional antipsychotic drugs in treatment-naive first-episode schizophrenia: a 52-week randomized trial of clozapine vs chlorpromazine. Neuropsychopharmacology 2003; 28(5):995–1003.
  • 3. Hendrich AB, Lichacz K, Burek B, et al. Thioridazine induces erythrocyte stomatocytosis due to interactions with negatively charged lipids. Cell Mol Biol Lett 2002; 7:1081–6.
  • 4. Pickholz M, Oliveira Jr ON, Skaf MS. Interactions of chlorpromazine with phospholipid monolayers: effects of the ionization state of the drug. Biophys Chem 2007; 125(2-3):425–34.
  • 5. Lieber MR, Lange Y, Weinstein RS, et al. Interaction of chlorpromazine with the human erythrocyte membrane. J Biol Chem 1984; 259(14):9225–34.
  • 6. Ivanov IT, Parvanova BK. Thermal dielectroscopy study on the vertical and horizontal interactions in erythrocyte submembrane skeleton. Electrochimica Acta 2019; 317:289–300.
  • 7. Ivanov IT, Parvanova BK. Differential dielectroscopic data on the relation of erythrocyte membrane skeleton to erythrocyte deformability and flicker. Eur Biophys J 2021; 50:69–86.
  • 8. Blanc L, Salomao M, Guo X, et al. Erythrocyte membrane-skeletal cohesion and its control of spectrin self-association. Biochemistry 2010; 49:4516–23.
login to comment