Photophysical characterization of sickle cell disease hemoglobin by multi-photon microscopy

: The photophysical properties of human sickle cell disease (SCD) Hemoglobin (Hb) is characterized by multi-photon microscopy (MPM). The intrinsic two-photon excited fluorescence (TPEF) signal associated with extracted hemoglobin was investigated and the solidified SCD variant (HbS) was found to demonstrate broad emission peaking around 510 nm when excited at 800 nm. MPM is used to dynamically induce and image HbS gelling by photolysis of deoxygenated HbS. For comparison, photolysis conditions were applied to a healthy variant of human hemoglobin (HbA) and found to remain in solution not forming fibers. The use of this signal to study the mechanism of HbS polymerization associated with the sickling of SCD erythrocytes is discussed.


Introduction
Sickle cell disease (SCD), the most common hereditary blood disorder worldwide, causes a number of health complications including chronic pain, anemia, chronic infection and stroke. It is caused by a single point mutation resulting in an amino acid substitution in the oxygen transport molecule hemoglobin (Hb). It has been thought for decades that the polymerization of HbS into rigid fibers is the primary pathogenic event causing the deformation of red blood cells [1][2][3]. These deformed, or sickled, red blood cells (RBC) have shortened circulation life, irregular surface and rheological properties, are mechanically inflexible, and are thought to contribute to clotting [4][5][6][7]. Nearly 100,000 Americans and millions worldwide, suffer from SCD and as such, are at risk of early mortality [8]. Though the disease has been known for over a century few safe and effective treatment options are available. Various environmental and treatment conditions, such as hydration, transfusions, fetal hemoglobin, and hydroxyurea are known to impact the rate and degree of sickling and therefore vaso-occlusive crisis [5,9,10], only limited success on different patients has been found. Therefore, it is of value to develop methods in which the effects of experimental treatments could be systematically studied in a controlled and clinically relevant environments. Cytological and rheological methods have been demonstrated to study macroscopic effects of various environmental and treatment conditions on the sickling rate [9,11] in addition to older, direct microscopic visualization utilized to study the more fundamental, sub-cellular kinetics of the sickling mechanism [4,12]. To compliment microscopic investigations of SCD hemoglobin, MPM is proposed as a means of direct visualization of the pathogenic sickling event by utilizing the intrinsic TPEF signature described here. TPEF emission for Hb has been observed elsewhere [13][14][15] and is used here for the direct visualization and stimulation of SCD Hb fiber formation. Studying SCD sickling kinetics by fluorescence microscopy is particularly interesting because simultaneous measurements of various environmental factors, such as dissolved oxygen [16], pH, concentration of labeled treatment molecules or temperature, could be acquired alongside those of HbS, therefore, enabling a systematic approach of screening interventions which removes some of the inherent ambiguity of the clinical environment. In order for intrinsic emission of sickled HbS to be utilized in such a fashion, the TPEF spectra of sickled HbS is characterized and presented. Using a deoxygenation agent and a heating or photolysis method to induce gelation, the sickling event, captured by labelfree MPM, is also demonstrated. For comparison a normal and healthy variant of hemoglobin, HbA, was exposed to the same conditions and shown not to form fluorescent fibers. Also demonstrated is the formation of structured HbS fibers in the presence of local impurities and externally applied magnetic fields. The use of MPM as an investigative research tool to further illuminate environmental conditions impacting the dynamics of HbS polymerization is discussed.

Imaging and spectral characterization experimental details
MPM imaging was performed on a Nikon A1R-MP Confocal system using a 40x, 1.3 NA, oil-immersion objective detecting 3-channels (Blue: 482/35, Green: 525/50, Red: 575/50) by episcopic GaAsP PMTs. Excitation was generated by femtosecond pulsed Mai Tai HP DeepSee laser. An IR stop filter (680 nm short-pass, OD 8 + ) was used to pass TPEF photons and reject reflected/scattered excitation light. Samples were excited at λ = 800 nm corresponding to an apparent maxima due to both optical collection efficiency for this particular microscope and TPEF cross section reported elsewhere [13]. Laser power level was measured at the sample plane using a PM100D digital optical power meter and S120C detector (Thorlabs Inc., Newton, New Jersey) Time lapse, 3-channel imaging was performed with the following set of video collection parameters -frame rate: 1/8 fps, 512x512 px (0.62 µm/px) frame size, equal detector gain and offset settings for each channel. To image HbS fibers gelled by external heating methods similar to those found in [3,17] moderate power settings of tens of mW were employed. To acquire spectral profile or perform deoxygenation by photolysis for time lapsed imaging, higher power, 30-60 mW, were needed since direct deoxygenation of Hb was employed instead of carbon monoxide lysis [2][3][4]. Two photon emission analysis was carried out first because of the high power required to perform simultaneous photolysis and imaging. Secondly, two-photon is favorable compared to single photon because the option to investigate HbS gel formation in thick scattering samples. Spectrum analysis was performed using 32-channel PMT collecting from 450 to 650 nm at 10 nm wavelength resolution. Spectral characterizations of HbS at long exposure durations at the high excitation powers were performed to ensure the dynamic processes of photolysis induced gelation were performed at doses below that to cause significant photo damage. Post processing was performed on images and data after acquisition via MATLAB (Mathworks, Natick, MA) and ImageJ (NIMH, Bethesda, MD).

Gelation
Samples were heated by hot plate or placed under the MPM and excited by IR laser pulses to induce gel formation. Neodymium magnets of dimensions 3/4" x 3/8" x 1/8" thick with a through thickness field of 0.28 T were placed across samples to study the structures formed by the diamagnetic sickled HbS fibers. Gelation was first confirmed macroscopically by the transition from clear to cloudy in a thin sample chamber of concentration HbS. HbA as a negative control was also tested in this way to rule out other mechanisms, such as burning or dehydration, as the ultimate source of the signal.

Intrinsic emission of HbS
HbA samples, either heated externally or photo disassociated by MPM excitation, did not demonstrate the formation of fibers, though low level fluorescence was detected as seen in Fig. 1(A). This was similar to the low level fluorescence observed in HbS samples before high excitation conditions were applied, suggesting that the fluorescence is a result of the general Hb structure and not wholly specific to HbS. HbS, prepared in high concentrations and heated or photo disassociated by high MPM excitation, was observed to form fibril structures which fluoresced significantly as imaged in Fig. 1(B). The emission spectra of HbS fibers was acquired and is shown in Fig. 2. Spectra represents the mean and standard deviation of 3 spectral measurements when excited at 800 nm and shows an emission peak around 510 nm. Dashed line is included as a guide for the eye. The origin of the signal may be due to a weak TPEF present in general Hb structures, however, upon the formation of the HbS gel, HbS molecules become organized in local high concentration fibers thereby increasing the apparent signal to measureable levels. Since only low level florescence is detected before HbS polymerization and two-photon laser excitation induces the polymerization process, it is difficult to quantify local concentration. Additionally it is difficult to measure TPE cross section of the dynamically changing environment as more fibrils begin forming. However, we have observed a relatively flat power-corrected florescence emission from HbS between 750 and 850 nm excitation Given the intrinsic HbS emission signal described here, MPM could be used to study the dynamics of HbS sickling over time with the application of heat or light to lyse bound oxygen from the HbS molecule.

Sickling event
MPM was used to investigate the dynamics of the sickling event in low and high concentrations of HbS. Figure 3 shows stills from the time evolution of HbS gelation in a 100 mg/ml and 250mg/ml solution of HbS. Over time, it is seen that the amount fluorescence's increases as HbS falls out of solution forming solid HbS structures. HbS in the low concentration state, such as observed in Fig. 3(A)-3(E), primarily formed into granulated clusters consistent with nucleation cites or spherulites observed in other studies [4,12]. In contrast, at high concentrations, Fig. 3(F)-3(J), HbS tended to form long fibrous structure consistent with the structures proposed to be responsible for the deformation of red blood cells [1,3,18,19]. A possible explanation for the resulting differences in structure is the lack or relative abundance of material in the local environment causing a different structure, i.e. that more hemoglobin molecules may be required for form long fibrils. Additionally, it was found that at low concentration, HbS forms micro cluster rapidly, but the growth of these clusters is slow compared to the rapid and complete growth of fibers in high concentration HbS samples. This may be due to the low availability of HbS molecules present in solution after initial cluster formation and limited by diffusion. These results show how different simulated blood cell environments may impact HbS sickling mechanisms as a function of HbS concentration.
Additionally, the structure of HbS formation in the presence of a magnetic field and sample impurities was studied. It is known, due to the ferrous components of Hb, that SCD red blood cells will align perpendicularly to the applied field when sickled [1], [20]. To demonstrate this phenomenon and further confirm the structure is in fact due to a systematic organization of HbS molecules instead of some unknown arrangement of non-ferrous proteins, the gelation of HbS was performed in the presence of a horizontally and vertically oriented magnetic fields, such as seen in Fig. 4(C) and 4(E). A still from the horizontally applied field show fiber formation vertically, Fig. 4(D), while the still from the vertically applied field show fiber formation left to right, Fig. 4(F). HbS RBC align perpendicularly in an applied magnetic field and therefore it is expected that HbS, after polymerization, also aligns perpendicularly in a magnetic field.
Both horizontal and vertical fields are used ensure the alignment is due to the presence of a strong magnetic field as opposed to some unknown local or system condition. Samples in the presence of impurities, such as bubbles or dust particles, show fiber formation is impacted by local structure as seen in Fig. 4(B) as fibers grow around air bubbles. Fig. 4. HbS fibers form with structure guided by surrounding structure or in the presence of magnetic fields. In a thin preparation (A) containing impurities, such as air bubbles, HbS fibers form around the impurities (B). In the presence of a magnetic field, the deoxygenated HbS fibers align perpendicularly to the applied field similar to the alignment of SCD erythrocytes elsewhere due to the diamagnetic nature of the molecule [1]. C and E diagram the orientation of the sample with respect to the horizontally and vertically applied fields respectively. D and F are the MPM images of fibers formed in the applied fields. Video content for B available as part of a 6540 kb Visualization 3, for D as a 1400 kb and for F as a 1501 Visualization 4 in supplemental material.

Discussion
Here, the intrinsic two-photon emission of solidified HbS was measured. This work provides the proof of concept for studying the dynamic mechanisms of SCD HbS sickling using intrinsic two-photon emission. The origin of the signal may be due to a weak TPEF present in general Hb structures, however, upon the formation of the HbS gel, HbS molecules become organized in local high concentration fibers thereby increasing the apparent signal to measureable levels. Also demonstrated here is the absence of sickling event found in HbA samples and evidence concerning magnetic field and sample impurities effects on the specific structures formed by sickling HbS. This work has not fully characterized the absorption cross section of HbS nor has it confirmed with structural analysis the fibers formed are in fact similar to those arrangements responsible for RBC deformation in SCD. However, this work has demonstrated a broad emission characteristic of HbS that can be simultaneously detected and stimulated using two-photon microscopy. Additionally, the use of magnetic fields in the work has confirmed the structures formed do behave as fibers of organized HbS is expected to [1,3,19].
This signal may also be present in SCD blood cells and could be employed as a biomarker of sickled erythrocytes or blockages in small vasculature in SCD affected tissue, such as the brain. This work is the initial characterization of this intrinsic signal. While the absolute cross section of HbS as a fluorescence molecule was not characterized here, the relative brightness of HbS excited even at modest powers after gelation, was found to be measureable and therefore useful for future investigations.
Also presented here are the differing structure which HbS takes on upon gelation when prepared in low and high concentrations. This observation affirms variation in severity of sickling-related SCD symptoms is related to hydration levels. This systematic method of study could be used to investigate critical levels of hydration. Also using this technique, studies concerning mixtures of Hb, such as those found in heterozygous individuals and patients treated with fetal hemoglobin, could be performed. Coupling this intrinsic signal with oxygen sensing measurements, would allow additional insight into the environment in which sickling becomes common.

Conclusions and future work
The intrinsic two-photon fluorescence spectrum of gelled HbS has been presented. The emission spectra was characterized. Gelation was carried out by application of light to perform oxygen lysing and induce sickling events. HbS structure formation was observed dynamically by measuring the intrinsic emission signal and found to form different structures depending on concentration. Fiber formation was also found to be influenced by sample defects, such as bubbles, and externally applied magnetic fields. HbA as a negative control was found to exhibit low level auto fluorescence, however did not form organized structures upon deoxygenation. The intensity of HbS as a fluorophore was found to be easily measureable ex vivo and therefore provides and intrinsic optical biomarker for the study of SCD. Although the signal is not suspected to be bright or spectrally distinct for in vivo diagnostics, this platform enables the study of label-free SCD sickling dynamics Future work will focus on TPEF emission cross section characterization the HbS fibers and could lead to evaluation of new therapies for SCD. Different conditions, such as treatments and oxygen levels, on the rate or degree of gelation could be investigated using this technique to further illuminate the mechanisms and factors contributing to severe SCD symptoms. Future studies will incorporate measurements of dissolved oxygen and other environmental conditions to understand how each impact sickling events.