Laser-induced primary and secondary hemostasis dynamics and mechanisms in relation to selective photothermolysis of port wine stains

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.


Introduction
Selective photothermolysis (SP) is a standard treatment modality for superficial vascular anomalies such as port wine stains (PWS) and other vessel-related dermatological disorders [1]. SP is based on the conversion of radiant energy to heat by (de)oxyhemoglobin and the thermal denaturation of blood and vascular tissue as a result of heat diffusion, referred to as photocoagulation or the photothermal response. By employing an appropriate wavelength and irradiance, supracritical temperatures (>70 8C) can be generated in the vessel lumen and confined spatially when the pulse duration is shorter than the time required for cooling of the target structures [2], i.e., 0.5-10 ms for PWS vessels of 30-300 mm in diameter [3,4]. Normal-sized capillaries and post-capillary venules (4-26 mm inner diameter [5]) have relatively short thermal relaxation times and thus remain spared during longer pulse durations, inasmuch as heat diffusion from these vessels precludes the generation of supracritical temperatures.
Background: Superficial vascular anomalies such as port wine stains are commonly treated by selective photothermolysis (SP). The endovascular laser-tissue interactions underlying SP are governed by a photothermal response (thermocoagulation of blood) and a hemodynamic response (thrombosis). Currently it is not known whether the hemodynamic response encompasses both primary and secondary hemostasis, which platelet receptors are involved, and what the SP-induced thrombosis kinetics are in low-flow venules. Objectives: To (1) define the role and kinetics of primary and secondary hemostasis in laser-induced thrombus formation and (2) determine which key platelet surface receptors are involved in the hemodynamic response. Methods: 532-nm laser-irradiated hamster dorsal skin fold venules were studied by intravital fluorescence microscopy following fluorescent labeling of platelets with 5(6)-carboxyfluorescein. Heparin and fluorescently labeled anti-glycoprotein Ib-a (GPIba) and anti-P-selectin antibodies were administered to investigate the role of coagulation and platelet receptors, respectively. Lesional sizes were quantified by software. Results: Laser irradiation consistently produced sub-occlusive thermal coagula. Thrombosis was triggered in all irradiated venules in a thermal coagulum-independent manner and peaked at 6.25 min post-irradiation. Heparin decreased the maximum thrombus size and caused thrombosis to reach a maximum at 1.25 min. Immunoblocking of GPIba abated the extent of thrombosis, whereas immunoblocking of P-selectin had no effect. Conclusions: The hemodynamic response ensues the photothermal response in a thermal coagulumindependent manner and involves primary and secondary hemostasis. Primary hemostasis is mediated by constitutively expressed GPIba but not by activation-dependent P-selectin.
The therapeutic efficacy of SP with respect to PWS depends on the extent to which the target vasculature is afflicted by the photothermal response. Complete photocoagulation/occlusion of the vascular lumen is required for optimal lesional clearance [6], which occurs in approximately 40% of patients [7]. Contrastingly, moderately responding (20-46%) and recalcitrant (14-40%) PWS [7,8] exhibit varying degrees of incompletely photocoagulated vessels containing semi-obstructive thermal coagula ( Fig. 1a-d). Consequently, alternative treatment strategies are needed to improve therapeutic outcomes in the poorly-to-non-responsive PWS patient population [9].
Recently an experimental modality, site-specific pharmacolaser therapy (SSPLT) [10], was proposed to ameliorate PWS recalcitrance to conventional SP. The conceptual framework of SSPLT is based on the occurrence of a hemodynamic response (thrombosis) following the photothermal response [11][12][13]. Accordingly, SP is combined with the prior administration of a prothrombotic-and/or antifibrinolytic-containing drug delivery system to instill complete occlusion, and thus complete removal, of semi-photocoagulated target vessels by the pharmacological augmentation of the hemodynamic response ( Fig. 1e-h).
T h ea i m so ft h es t u d yw e r et od e t e r m i n et h er o l ea n dt h e kinetics of primary and secondary hemostasis in the hemodynamic response and to establish which key platelet surface receptors are involved in SP-induced thrombosis using a hamster dorsal skin fold model in conjunction with fluorescent labeling of platelets and intravital fluorescence microscopy. The findings comprise a basis for the continued development of SSPLT in an effort to optimize the laser treatment of aberrant cutaneous vasculature.
Online supplemental material is indicated with a prefix 'S' throughout the manuscript.

Animals
The animal protocol was approved by the Lille University Hospital animal ethics committee. Animals were treated in compliance with NIH publication 86-23.
Sixty two male Golden Syrian hamsters (93-117 g, Dé pré , Saint Doulchard, France) were anesthetized by intramuscular injection of ketamine (200 mg/kg), xylazine (10 mg/kg), and buprenorfine (0.02 mg/kg) after pre-anesthesia with diethyl ether. The dorsal chamber, which has been employed previously as a preclinical Cartoon illustrating the principles of selective photothermolysis (SP) vs. site-specific pharmaco-laser therapy (SSPLT). Irradiation of port wine stain (PWS) vessels with a yellow laser (a) is characterized by the photothermal response (b), resulting in the formation of a thermal coagulum (b, insert), and the hemodynamic response, which is characterized by thrombosis. In time the thrombus deteriorates (c), leaving a partially occluded lumen (insert: cross-sectional view: 1, thermal coagulum; 2, thrombus; 3, patent lumen). In (d), a histological section imaged by scanning electron microscopy is presented of a PWS blood vessel following 577-nm laser irradiation, exhibiting a similar damage profile as in (c) (scale bar = 10 mm). The vessel is partially occluded by a thermal coagulum (TC) composed of electron-dense, thermally denatured blood with numerous electron-lucent spherical structures (insert upper left, scale bar = 1 mm) and a thrombus (T) comprising aggregated platelets and associated fibrin. The rest of the lumen is patent (P) and contains intact erythrocytes (IE). Endothelial cells in the thermally afflicted region exhibit cytoplasmic vacuolisation and separation from the basal lamina (insert bottom right, arrow, scale bar = 1 mm). Image adapted from Ref. [10]. SSPLT is an envisaged treatment modality for refractory PWS, whereby SP is combined with the systemic administration of a prothrombotic-and/or antifibrinolytic-containing drug carrier (e). It is aspired that, following conventional SP, the drug carrier accumulates in the thrombus (f) and its contents are released by a second stimulus (e.g., heat generated by a near-infrared laser pulse) (g), resulting in local hyperthrombosis and inhibition of fibrinolysis and complete occlusion of the vascular lumen (h). Clinical evidence suggests that complete occlusion of target vasculature (h, insert: 1, thermal coagulum; 2, thrombus) is associated with well-responding lesions. model for PWS vasculature [15][16][17], was implanted according to Bezemer et al. [15]. Reagents were infused through the subclavian vein with a 30 G needle ($200 mL/30 s). The animals were sacrificed by intravenous administration of potassium chloride.
Raising the temperature above the phase transition temperature of the lipid bilayer (55.5 8C, Fig. S1) is associated with a gel-toliquid-crystalline phase transition of component phospholipids, leading to rapid release of CF and a reduction in fluorescence quenching. Hyperfluorescence during intravital microscopy following the laser pulse therefore served as an indicator for thermal denaturation of plasma components (>45 8C [19]) and erythrocytes (<51 8C [20]), i.e., precursor events in thermal coagulum formation.

Intravital microscopy, laser-induced thrombosis
The microscopy and thrombus induction setup is explained in Fig. 2a. To verify that non-obstructive thermal coagula ( Fig. 1c and d) were formed at the laser parameters employed for thrombus induction, the microscope optics were configured as described previously [15]. Vascular patency was determined by infusion of fluorescent microspheres (2 Â 10 9 microspheres/kg, 100-mLi njection volume) [15]. Thermal coagula (n = 15) were visualized using a filter set whose excitation wavelength range (l ex =540-580 nm) matches a high absorption range of thermally denatured blood [21] and whose emission range (l em = 605-655 nm) is broad enough to generate brightfield images in transillumination mode. In this configuration, simultaneous visualization of the laser-induced thermal coagulum (appearing dark against a lighter background), skin fold anatomy, and fluorescent microspheres (section 'Shear Rates') was possible. For thrombus imaging, mAband CF-labeled platelets were visualized at l ex =480AE 15 nm and l em =535AE 20 nm. In some antibody experiments, transillumination was applied for contrast enhancement. Endovascular events were recorded for a period of 8 min in the thermal coagulum experiments and for 15 min in the thrombosis experiments.

Intravital platelet labeling
After implantation of the dorsal skin fold chamber, platelets were stained cytochemically by infusion of 180 mL of the liposome suspension (n = 17) to study the involvement of primary hemostasis. Resting and activated platelets endocytose free CF [22],i n f u s e di n unencapsulated form and released from the liposomes, resulting in the formation of a fluorescent thrombus following laser-induced endovascular damage. High molecular weight heparin was administered after liposome infusion at a concentration of 2000 IU/kg (n = 11) to study the involvement of secondary hemostasis.  (1) and quantified for pixel area (A pix ) and total intensity (I tot ) using dedicated software. For the analysis of antibody-stained thrombi (bottom panel), the vascular lumen was roughly contoured and saved as a separate image file (2). An intensity histogram of background (2, encircled) was collected from frames of each video sequence at t = 0, 2.5, 5.0, 7.5, and 10.0 min. To characterize the thrombus (3), a threshold range was defined in the programmed macro that had an intensity cut-on value of 5 units higher than the highest recorded intensity value in the 5 background frames. The thrombus was quantified for pixel area (A pix , the total number of 'thrombus pixels') and total intensity (I tot , cumulative intensity of the 'thrombus pixels'). Lesional size was calculated by normalized (A pix ) Á normalized (I tot ) and relative lesional size was calculated by [normalized ( Activated platelets were stained by infusion of CD62P-Alexa488 mAbs (500 mg/kg, n = 10) to assess the role of P-selectin in the hemodynamic response. IgG-FITC (500 mg/kg, n =4) was used as isotype control to determine the extent of non-specific binding. The anti-CD62P mAbs exhibited cross-reactivity with activated but not resting hamster platelets (section-S4), although some nonspecific binding to resting hamster platelets was observed (Fig. S2). The anti-CD62P mAbs induced negligible reduction in platelet count (section-S6). All mAbs were diluted with 0.9% sodium chloride solution to a final injection volume of 200 mL.

Laser irradiation
The target venules had a mean AE SD diameter of 157 AE 35 mm (range = 86-252 mm). Lesions were induced with a frequencydoubled Nd:YAG laser (532 nm, Entertainer, Laser Quantum, Cheshire, UK) at a power of 224 mW, a mean AE SD incident radiant exposure of 289 AE 38 J/cm 2 , a 2.3 Â 10 À3 -mm 2 spot size, and a 30-ms pulse duration [15]. The pulse duration for this wavelength falls within the clinically employed range [23].

Shear rates
Blood flow measurements were performed using fluorescent microspheres. Shear rates were determined (n = 95) based on the measured blood flow velocity and vessel radius (section-S7). At a mean AE SD flow of 0.48 AE 0.21 mm/s and a mean AE SD venular diameter of 0.15 AE 0.03 mm, the mean AE SD shear rate was 7.0 AE 3.7 s À1 (range = 1.8-19.6 s À1 ).

Lesion quantification
The quantification of laser-induced lesions was performed differently for cytochemically (CF-) vs. immunolabeled platelets (Fig. 2b, section-S8). For the former, isolated video frames of laserinduced lesions were manually contoured and quantified for pixel area (A pix ) and total intensity (I tot ) in SigmaScan Pro (Systat Software, Point Richmond, CA). Contrastingly, immunolabeled thrombi were demarcated in SigmaScan Pro using a thresholding algorithm, whereby 'thrombus pixels' were defined as pixels with an intensity of 5 grayscale units above the highest background intensity, and quantified for A pix and I tot . For both quantification techniques, A pix was normalized to baseline (t = 0, lesional area immediately after the laser pulse) and I tot was normalized to the maximum total intensity. Lesional size was defined by normalized (A pix )Ánormalized (I tot ) in both quantification methods and expressed comprehensively (thermal coagulum + thrombus) or individually (thrombus). Normalization of A pix and I tot and lesional size calculations were performed per experiment/animal. Lesional sizes were averaged per experimental group, normalized to baseline, and plotted as a function of time. The relative lesional size was calculated by dividing the lesional size at each time interval by the baseline value.
Additionally, a computational analysis method was developed for validating the manual image analysis. Good agreement was found between manual and automated analysis (section-S9).

Statistical analysis
Statistical analysis (means, standard deviations, and independent hetero-and homoscedastic Student's t-tests) was performed with SPSS (SPSS, Chicago, IL). Shapiro-Wilk tests confirmed the normal distribution of continuous data. The type of t-test used was based on Levene's test of equality of variances. A p-value of 0.05, designated by (*) throughout the text, was considered statistically significant. A p-value of 0.01 is designated by (**).

Induction of subocclusive thermal coagula
The laser-induced formation of subocclusive thermal coagula (photothermal response) was investigated first to ensure that our model emulated the endovascular damage profile associated with SP-treated refractory PWS vasculature (Fig. 1d). Laser irradiation consistently resulted in the formation of subocclusive thermal coagula in all irradiated vessels as evidenced by the uninterrupted flow of fluorescent microspheres (Video-S1/S2). Thermal coagula remained attached to the vessel wall ( Fig. 3a-d) or detached within a few seconds after the laser pulse ( Fig. 3e-h).

The photothermal response triggers primary and secondary hemostasis
Next, the manifestation of primary hemostasis (platelet aggregation) following photothermolysis was studied. Platelets were labeled by the systemic infusion of CF and liposomes containing a self-quenching concentration of CF were co-infused to serve as a thermal damage indicator. Laser irradiation was associated with transient hyperfluorescence as a result of a heat-induced liposomal membrane transition and release of CF (Fig. 4b, arrowhead, Video-S3), confirming local denaturation of plasma proteins and erythrolysis (thermal coagulum formation).
Platelet adhesion and the development of a nidus occurred in all irradiated venules within seconds after thermal coagulum formation ( Fig. 4a-f), which was characterized by a rapid growth phase during the first 1.25 min and a slow growth phase in the subsequent 5.0 min. At 6.25 min the comprehensive lesional size reached a maximum with a 9.5-fold greater volume with respect to baseline (**, Fig. 4m and o). The slow growth phase encompassed extensive thromboembolic activity in which clot build-up exceeded breakdown, given the zigzag pattern of the maxima and the upward trend of the lesional size curve, respectively (Fig. 4m). The slow growth phase was ensued by clot breakdown as evidenced by the rapid decline in lesional size after 6.25 min (Fig. 4m and o) and extensive embolization of large platelet aggregates (Video-S3). At 15.0 min, the lesions had gradually sloughed off to 58% of their maximum (**). Platelet adhesion and aggregation at the site of laser irradiation occurred in the presence and absence of a thermal coagulum (Fig. 5a-f).

[ ( F i g . _ 5 ) T D $ F I G ]
Heparin, an antagonist of factor (f)Xa and thrombin formation [24], was infused to determine the involvement of coagulation (secondary hemostasis) (Fig. 4g-l). Heparin reduced the maximum lesional size to 49% of the heparin-untreated group (p = 0.07, Fig. 4n vs. m) and caused the peak of thrombosis to occur at 1.25 min (Fig. 4n and o). At this time point the comprehensive lesional size had increased 2.6-fold (Fig. 4o), whereby the maximum relative lesional size was 27% of that in the heparinuntreated group (*).
In contrast to the heparin-untreated group, clot breakdown started after 1.25 min (Fig. 4n) and encompassed extensive embolization of platelet aggregates and, in some instances, portions of the thermal coagulum. The breakdown phase plateaued at 6.25 min at a lesional size that did not differ from baseline (p = 0.08), suggesting that the presence of residual platelet aggregates was minimal and/or that the thermal coagulum had partly or entirely dissociated from the vascular wall. As reported for the heparin-untreated group, the presence of a thermal coagulum was not required for platelet adhesion and aggregation to occur at the site of laser irradiation (Fig. 5g-l).
The accumulation of CF-labeled platelets at the site of endovascular damage and the inhibitory effect of heparin demonstrate that the photothermal response triggers primary as well as secondary hemostasis, respectively, in a thermal-coagulum adhesion/embolization-independent manner. To examine which key platelet receptors are involved in cell adhesion/aggregation during the hemodynamic response, the hemodynamic response was studied following immunoblocking of GPIba (CD42b) and Pselectin (CD62P).

Partial inhibition of CD42b reduces the extent of laser-induced thrombosis
CD42b is a transmembrane subunit of the constitutively expressed GPIb-IX-V receptor complex with heterotypic binding sites for von Willebrand factor (vWF), Mac-1, CD62P, athrombin, fXI/XIIa, and high-molecular-weight kininogen [25]. Although the major physiological function of CD42b is the adhesion of circulating platelets to vWF in the subendothelial matrix at high shear, which leads to activation of integrin a IIb b III (GPIIb/IIIa) and subsequent aggregation [26], it has also been shown to mediate platelet adhesion under low shear conditions, i.e., in venules [27,28]. Inhibition of CD42b has further been correlated to significantly reduced platelet microparticle formation [29]. At a shear rate of 7.0 AE 3.7 s À1 in hamster venules, CD42b may therefore constitute an important receptor during the hemodynamic response, inasmuch as both platelet aggregation and coagulation prevail.
Anti-CD42b mAbs were infused to investigate the role of GPIba in the hemodynamic response (Fig. 6a-f). Fractional immunoblocking of CD42b imposed no deleterious effect on thrombus formation during the first 3.5 min compared to CF-stained lesions, but significantly reduced clot size during the subsequent 8.0 min (*, Fig. 6m). Thrombosis peaked at 2.75 min (Fig. 6ma n do ) followed by a steep deterioration phase, characterized by embolization of platelet aggregates and, sparsely, portions of the thermal coagulum, that stabilized at 4.25 min. Thrombi remained enveloped around the thermal coagulum up to 15.0 min (** vs. baseline, p = 0.09 vs. CF), suggesting that, at an estimated 3% GPIba inhibition, clot fortification by cross-polymerized fibrin imposed greater resistance to deterioration than a clot composed of predominantly platelets (Fig. 6m, CD42b vs. CF + Hep).

Inhibition of CD62P does not affect the extent of laser-induced thrombosis
CD62P is a cell adhesion molecule constitutively expressed in platelet a-granules [30] and endothelial cell Weibel-Palade bodies [31] that, upon cell activation, is translocated to the outer membrane to mediate platelet-platelet [32], platelet-leukocyte [33], and platelet-endothelial cell interactions [34]. In addition to cell recruitment, the expression of CD62P potentiates a procoagulant state by enhancing fibrin deposition [35] through the incorporation of P-selectin glycoprotein ligand (PSGL-1)-expressing, tissue factor-bearing microparticles derived from the abovementioned cells [36]. As CD42b, CD62P may therefore play an instrumental role in the laser-induced hemodynamic response.
Anti-CD62P mAbs were infused to assess the role of P-selectin in the hemodynamic response (Fig. 6g-l). With the known inhibitory properties of the RB40.34 clone (section-S4), it was expected that the anti-CD62P mAbs would reduce the extent of thrombosis. Although a slight reduction in thrombus size manifested itself in the rapid growth phase, no inhibitory effect was observed in the slow growth and breakdown phases vs. CFstained thrombi (Fig. 6n, black vs. dotted line, respectively). The maximum thrombus size was reached at 5.75 min and both lesional size curves (CD42b-and CF-stained thrombi) exhibited a similar progression up to 8.0 min. The 8.0-min time point marked a deflection in the downward trend in lesional size in the CD62P group, which may have been a result of increased CD62P expression (as evidenced by an increase in fluorescence intensity and not the lesional area, data not shown) and/or the nonspecific binding of the mAb (section-S4).

Discussion
A laser-mediated vascular injury model was employed with which the photothermal and hemodynamic response could be concomitantly studied in refractory PWS vessel analogues by intravital fluorescence microscopy. With this model we corroborated that primary hemostasis (platelet aggregation) comprises an integral component of endovascular laser-tissue interactions in semi-photocoagulated blood vessels [11][12][13]. Additionally, we demonstrated that (1) secondary hemostasis (blood coagulation) is triggered by SP, (2) primary and secondary hemostasis occur in the absence of a thermal coagulum, (3) CD42b is involved in the adhesion of platelets at the site of laser irradiation under low flow conditions, (4) the adhesion of platelets at the site of laser irradiation is accompanied by platelet activation, and (5) CD62P does not mediate platelet adhesion during the hemodynamic response. Moreover, the kinetics of laser-induced thrombosis in low-flow refractory PWS vessel analogues were established. Knowledge about the involved platelet receptor and thrombosis kinetics is particularly important in the continued development of SSPLT as a clinical modality.
As this study demonstrated, an intricate and causal relationship exists between the photothermal and the hemodynamic response. This relationship appears to be manifested at different levels: the nucleation centers (erythrocytes), the endothelium, and the thermal coagulum. Firstly, rapid heat build-up in erythrocytes causes the cells to swell and rupture [11,37]. Erythrocytes contain adenosine diphosphate [38] and phosphatidyl serine [39] that, when liberated and exposed, trigger platelet adhesion [40] and activation [41] and initiation of coagulation [42], respectively. Erythrocyte membrane disruption and complete disintegration have been reported to occur at 47-49 8C and 50 8C, respectively [20,37]. These temperatures were exceeded in the in vivo experiments as corroborated by the CF-encapsulating thermosensitive liposomes and by the consistent formation of thermal coagula at the employed laser settings. Furthermore, it has been shown that ruptured erythrocytes remain attached to the thermal coagulum [43], causing the biochemical template for thrombosis to become stationary following laser irradiation. Image analysis of laser-induced lesions confirmed that thrombo-embolic activity always occurred around the thermal coagulum, even when the thermal coagulum had attached to or was overlaying a nonirradiated portion of the vessel downstream of the irradiation site. Consequently, erythrolysis constitutes an important trigger of the hemodynamic response following the photothermal response. This is underscored by the finding that laser-induced thrombosis in vivo occurs in irradiated blood vessels perfused with washed erythrocytes, erythrocyte ghosts, or hemolysate but not with platelet-rich plasma, PBS, or a non-absorbing exogenous chromophore [12].
Secondly, the heat-afflicted endothelium evidently plays a role in the hemodynamic response. High-fluence laser irradiation of microvessels causes ultrastructural perturbation of the endothelial cell membrane and denudation of the endothelial monolayer [13,44,45], which triggers primary [46,47] and secondary hemostasis [42]. It should be noted that the volumetric heat production, m a Áf, where m a is the absorption coefficient (cm À1 ) and f is the fluence rate (J/cm 2 ), that led to such endothelial damage [13] was very high ($17,656 kJ/cm 3 , section-S10.1) and not representative for either the clinical setting or our model. However, Tan et al. [44,45], who employed a substantially shorter pulse duration and much larger spot size for treating PWS, showed that endothelial damage can be achieved at significantly lower m a Áf. Contrastingly, at a thousand-fold lower m a Áf of $16 kJ/cm 3 (section-S10.2), thermal coagulum formation and endothelial denudation were absent and only one component of the hemodynamic response (platelet adhesion and aggregation) prevailed, albeit transiently [18]. Inasmuch as our model produced a thermal coagulum at a m a Áf of $66 kJ/cm 3 at the blood-endothelium interface (section-S10.3), a contribution of thermally afflicted endothelium to the hemodynamic response is warranted. This was experimentally corroborated by the fact that thrombosis occurred in laserirradiated vessel segments where the thermal coagulum had dislodged (e.g., Fig. 3h, Video-S2) and is in agreement with previous reports [12].
Thirdly, photocoagulation is associated with protein denaturation [48,49] that embodies conformational rearrangements and cross-linking/aggregation of unfolded/misfolded proteins [50] and corollary amyloid fibril formation [51]. There is increasing evidence that misfolded proteins/amyloids activate platelets via CD42b [52] and initiate the contact activation pathway through the auto-activation of fXII by anionic surfaces [53]. Inasmuch as thermal coagula are in part comprised of thermally denatured (unfolded/misfolded) proteins and anionic moieties exposed on the surface of ruptured erythrocytes, these laser-induced lesions may constitute the basis for the initiation of primary and secondary hemostasis that persisted around the thermal coagula. These postulations are, however, hypothetical and are currently being investigated in a separate study.
Although several specific pathways underlying endovascular laser-tissue interactions require further elucidation, it is unequivocal that SP affects both primary and secondary hemostasis in semi-photocoagulated vasculature. Primary hemostasis involves platelet adhesion, activation, and aggregation. The immunostaining experiments with anti-CD42b and anti-CD62P mAbs revealed that platelet adhesion to the thermal coagulum and/or to the thermally afflicted vascular wall is mediated by GPIba, but not CD62P, and that the platelets become activated upon adhesion, respectively.
For platelet aggregation, integrin a IIb b III (CD41) activation is required [54], which could not be studied directly because none of the assayed antibodies cross-reacted with hamster a IIb b III .
Nevertheless, experimental data confirms the activation of integrin a IIb b III , and thus aggregation, in the hemodynamic response.
Strong evidence is provided by the facts that a >9-fold expansion of the thrombus following laser irradiation cannot be accounted for by solely platelet adhesion and that embolization of large platelet aggregates was observed in the breakdown phase. The involvement of GPIba in the adhesion process further implies activation of integrin a IIb b III (concurs with the binding of ligands to the GPIb-IX-V complex during platelet adhesion [55]), which was supported by the reduction in thrombus growth following partial immunoneutralization of GPIba. Also, a IIb b III activation was corroborated by the positive immunolabeling of CD62P that is co-expressed when a IIb b III is activated [56]. Finally, we performed the same experiments in the dorsal vasculature of 3 surplus BALB/c mice (where anti-CD41 antibodies do cross-react with the respective platelet epitope), and confirmed a IIb b III activation during laserinduced thrombosis (Video-S4).
During secondary hemostasis, prothrombin is converted to thrombin by the prothrombinase complex, which exerts pleiotropic effects on both platelets and coagulation. The binding of thrombin to the GPIb-IX-V complex induces platelet adhesion and spreading, dense granule secretion, and a IIb b III activation and subsequent aggregation [57]. It also accelerates the hydrolysis of protease-activated receptor (PAR)-1 (a thrombin receptor) [58] that further contributes to platelet activation. With respect to coagulation, thrombin mediates the cross-polymerization of fibrin that acts as a support matrix in thrombus fortification. The complete inhibition of thrombin generation at high heparin concentrations did not completely abrogate thrombus formation but substantially reduced the extent and duration of thrombosis. These observations confirm the involvement of the previously mentioned primary hemostasis mechanisms, which do not require thrombin, and attest to a role of coagulation in the hemodynamic response.
Inasmuch as thrombin generation is antagonized by heparin at the level of fXa in the common pathway, it is impossible to deduce whether coagulation was initiated through the tissue factor pathway or the contact activation pathway. The activation of coagulation in our model is, however, in contrast with previous studies [11,13,59], where no fibrin cross-polymerization (secondary hemostasis) was found following laser-induced thrombosis. It should be stressed, however, that the influence of secondary hemostasis was considerably inferior to that of primary hemostasis given that very mild inhibition of CD42b and complete inhibition of thrombin formation generated comparable thrombus kinetics curves. This is quite surprising given the fact that venous thrombosis predominantly entails activation of coagulation rather than platelet aggregation [60]. Unfortunately, the lack of mouse anti-human fibrin mAb cross-reactivity with hamster fibrin (section-S4) precluded the intravital qualification and quantification of fibrin cross-polymerization in the thrombi.
In conclusion, the hemodynamic response ensues the photothermal response and involves the activation of primary and, to a lesser extent, secondary hemostasis. Laser-induced venular thrombi reach a maximum size at 6.25 min after laser irradiation under very low flow conditions. Primary hemostasis, encompassing platelet adhesion, activation, and aggregation, is mediated by constitutively expressed GPIba but not by activation-dependent Pselectin.

Funding sources
This work was partially supported by the Technological Collaboration Grant (TSGE 1048) of the Dutch Ministry of Economic Affairs (MH, JFB), a grant from Novo Nordisk Farma BV (MH), and the Bloodomics project (6th Framework Program of the European Union (LSHM-CT-2004-503485)) (IIS, HD). A patent has been filed for the proposed technology (PCT 2010050833).