Histopathological analysis of zebrafish after introduction of non-biodegradable polyelectrolyte microcapsules into the circulatory system

Polyelectrolyte microcapsules are among the most promising carriers of various sensing substances for their application inside the bloodstream of vertebrates. The long-term effects of biodegradable microcapsules in mammals are relatively well studied, but this is not the case for non-biodegradable microcapsules, which may be even more generally applicable for physiological measurements. In the current study, we introduced non-biodegradable polyelectrolyte microcapsules coated with polyethylene glycol (PMs-PEG) into the circulatory system of zebrafish to assess their long-term effects on fish internal organs with histopathologic analysis. Implantation of PMs-PEG was not associated with the formation of microclots or thrombi in thin capillaries; thus, the applied microcapsules had a low aggregation capacity. The progression of the immune response to the implant depended on the time and the abundance of microparticles in the tissues. We showed that inflammation originated from recognition and internalization of PMs-PEG by phagocytes. These microcapsule-filled immune cells have been found to migrate through the intestinal wall into the lumen, demonstrating a possible mechanism for partial microparticle elimination from fish. The observed tissue immune response to PMs-PEG was local, without a systemic effect on the fish morphology. The most pronounced chronic severe inflammatory reaction was observed near the injection site in renal parenchyma and within the abdominal cavity since PMs-PEG were administered with kidney injection. Blood clots and granulomatosis were noted at the injection site but were not found in the kidneys outside the injection site. Single microcapsules brought by blood into distal organs did not have a noticeable effect on the surrounding tissues. The severity of noted pathologies of the gills was insufficient to affect respiration. No statistically significant alterations in hepatic morphology were revealed after PMs-PEG introduction into fish body. Overall, our data demonstrate that despite they are immunogenic, non-biodegradable PMs-PEG have low potential to cause systemic effects if applied in the minimal amount necessary for detection of fluorescent signal from the microcapsules.

Polyelectrolyte microcapsules are among the most promising carriers of various sensing substances for their application inside the bloodstream of vertebrates. The long-term effects of biodegradable microcapsules in mammals are relatively well studied, but this is not the case for non-biodegradable microcapsules, which may be even more generally applicable for physiological measurements. In the current study, we introduced nonbiodegradable polyelectrolyte microcapsules coated with polyethylene glycol (PMs-PEG) into the circulatory system of zebrafish to assess their long-term effects on fish internal organs with histopathologic analysis. Implantation of PMs-PEG was not associated with the formation of microclots or thrombi in thin capillaries; thus, the applied microcapsules had a low aggregation capacity. The progression of the immune response to the implant depended on the time and the abundance of microparticles in the tissues. We showed that inflammation originated from recognition and internalization of PMs-PEG by phagocytes.
These microcapsule-filled immune cells have been found to migrate through the intestinal wall into the lumen, demonstrating a possible mechanism for partial microparticle elimination from fish. The observed tissue immune response to PMs-PEG was local, without a systemic effect on the fish morphology. The most pronounced chronic severe inflammatory reaction was observed near the injection site in renal parenchyma and within the abdominal cavity since PMs-PEG were administered with kidney injection. Blood clots and granulomatosis were noted at the injection site but were not found in the kidneys outside the injection site. Single microcapsules brought by blood into distal organs did not have a noticeable effect on the surrounding tissues. The severity of noted pathologies of the gills was insufficient to affect respiration. No statistically significant alterations in hepatic morphology were revealed after PMs-PEG introduction into fish body. Overall, our data demonstrate that despite they are immunogenic, non-biodegradable PMs-PEG have low potential to cause systemic effects if applied in the minimal amount necessary for Introduction Injection of microcapsules into fish kidney 133 All experimental procedures were conducted in accordance with the EU Directive 134 2010/63/EU for animal experiments and the Declaration of Helsinki; the protocol of the study was 135 registered and approved before the start of the experiment by the Animal Subjects Research 136 Committee of the Institute of Biology at Irkutsk State University (Protocol №1/2017). 137 To deliver the microcapsules into the fish bloodstream, we injected the microcapsules 138 directly into the fish kidney, as described in detail elsewhere . Due to 139 the small size of D. rerio, it is difficult to administer PMs-PEG directly into the fish blood vessels; 140 therefore, injection into vascular-rich tissues to deliver microcapsules into the circulatory system 141 was performed as an alternative.

142
On the day of the experiment, the fish chosen from a random aquarium by lots were 143 individually transferred to a solution of an antiseptic (0.0002% methylene blue for 1.5 min), then 144 anesthetized in an emulsion of clove oil in water (0.1 ml/L) until the fish turns on its side and stops 145 responding to a light pinch of the fin (about 1 min). Individuals were immobilized on a wet sponge 146 and injected with 1.6 μl of microcapsules suspension in 0.9% NaCl (4.2*10 6 microcapsules per μl) 147 into a central bulge of the fish trunk kidney using a 31G (Ø 0.25 mm) needle connected to an IM-148 9B microinjector (Narishige, Japan).

149
To check the success of PMs-PEG delivery to the bloodstream, a rapid visual inspection of 150 the gills with a fluorescent microscope was made after cutting the fish gill cover (Borvinskaya et 151 al., 2018b). When the injection was performed correctly, fluorescent microcapsules occur in the 152 capillaries of the denuded fish gills; we aimed to obtain several dozens of PMs-PEG per gill, which 153 is enough to perform physiological measurements with a microencapsulated probe (Borvinskaya 154 et al., 2017). In the control group, the fish received intrarenal injection of saline. Finally, fish were 155 rinsed with water and returned to their holding bags. All staff conducting the experiment was aware 156 of the allocation of the fish groups during the experiment. The fish condition and mortality were 157 monitored twice a day.

Histological and histopathological examination
PeerJ reviewing PDF | (2020:11:55227:1:1:NEW 23 Mar 2021) Then mean recorded prevalence and affected area of viewed sections were multiplied on the 212 proportion of sections with pathology to take into account how much tissue was affected in the 213 third dimension. Thus, the equation for evaluation of final relative severity (S) of the particular 214 histopathology for each individual was: It should be noted that the evaluation of the severity and adversity of the morphological 224 alterations, which is usually challenging (Schafer et al., 2018), was not a direct purpose of this 225 study. Instead, we aimed to quantify the effect of microcapsules on the fish organism and directly 226 compare the obtained grades for morphological structures in fish treated with PMs-PEG and 227 injected with saline.

228
The obtained grades for individual morphological structures of the fish from the treatment 229 groups were compared with the nonparametric Mann-Whitney test with Benjamini-Hochberg 230 correction for multiple testing. Mulitple groups were compared using ANOVA with sex and post-231 injection day as factors. The analysis was performed with the coin (Hothorn et al., 2008), magrittr 232 (Schmidt, 2019) and base stats packages for R (R Core Team, 2017). Manuscript to be reviewed Fluorescent PMs-PEG were implanted into the circulatory system of adult zebrafish D. rerio 239 by intrarenal injection. The highest mortality (5-6% in both experimental and control groups) 240 occurred immediately after the injection due to the stress and injury caused by the injection. Those 241 fish that woke up after injection with saline did not die during the experiment. The cumulative 242 mortality of fish that received an intrarenal injection of PMs-PEG during the first week was 18%. 243 Further presence of microcapsules in the fish did not cause noticeable effects at the organismal 244 level, thus suggesting a low level of direct toxicity of PMs-PEG. This outcome is consistent with 245 those obtained earlier for fish received an intrarenal injection of 1.6-5 µm diameter PMs-PEG 246 (Borvinskaya et al., 2018a).

247
The microstructure of the kidneys, liver, and gills of fish was monitored for 22 days after the 248 injections to clarify interaction of PMs-PEG with fish tissues. During an injection into the fish 249 kidney, most of the microcapsules spill into the abdominal cavity. Therefore, the largest number 250 of microcapsules on histological sections can be seen around the swimming bladder, at the bottom 251 of the abdominal cavity and between the internal organs. Moreover, a large concentration of 252 microcapsules is present in the renal parenchyma at the injection site.

253
The microcapsules that had entered blood vessels (due to their rupture with a needle) were 254 detected in the peripheral parts of the kidney one day after injection, as well as in the liver and 255 gills (Fig. 1, Supplementary Table 2). First day observations show that only a minimum amount 256 of microcapsules entered the skeletal muscles and brains of the fish. These histological 257 observations are generally consistent with the previous external examinations of fish organs after 258 the PMs-PEG intrarenal administration, which demonstrated immediate distribution of 259 microcapsules mainly in the liver, gills, kidney, and less often in the muscles, gonads, and fin 260 capillaries (Borvinskaya et al., 2018a).

261
Two weeks after injection, histological examination revealed that the microcapsules were 262 still present in the kidneys and liver of the fish, supporting the observation of PMs-PEG 263 accumulation in organs with a rich capillary network. In the second week after injection, 264 microcapsules also appeared in the intestinal epithelium and single PMs-PEG were found in the 265 spinal cord and eyes. In the muscles and brain, the microcapsules were still located, albeit rarely 266 and singly.

276
In the third week, the microcapsules were still common in the kidneys and liver, and were 277 also abundant in the gills, intestine and muscles of fish (Fig. 1). Separate microcapsules and PeerJ reviewing PDF | (2020:11:55227:1:1:NEW 23 Mar 2021) Manuscript to be reviewed 278 aggregates were found in skeletal muscles, in the brain and spinal canal. Except for one individual 279 with plenty of PMs-PEGs observed in the heart ( Fig. 2A), in other fish only rare microcapsules 280 were found in central vessels (Fig. 2).

287
We did not observe an increase in the number of PMs-PEGs in the liver, kidneys and gills 288 of fish during the three weeks after injection (Supplementary Table 2), suggesting that the majority 289 of microcapsules carried with the blood throughout the fish body quickly settles in these organs as 290 what has been previously shown in mouse models (Sindeeva et al., 2020). However, at least some 291 PMs-PEG were able to migrate from the tissues where they were originally settled and 292 accumulated in the skeletal muscles and intestine.

293
The histological examination of the D. rerio intestine showed that engulfed PMs-PEG were 294 carried into the intestinal mucosa by phagocytic cells (Fig. 3). These loaded phagocytes tend to 295 gather in the intestinal epithelium, forming cavities between enterocytes, which can rupture 296 ( Figure 3D) and release their contents into the intestinal lumen ( Figure 3C). Thus, fluorescent 297 particles enter the intestinal lumen (Fig. 3C, 3D) and can be excreted in feces. This indicates that 298 the fish can at least partially get rid of the irritating agent implanted in the body. However, the 299 existence and effectiveness of such a mechanism of microparticle excretion have almost never 300 been studied.

301
It is known that particles larger than 0.5 µm cannot pass through the epithelium of blood 302 vessels and thus cannot be removed through the excretory organs (De Jong et al., 2008). Therefore, 303 they usually remain in the body and cause chronic inflammation, which at best results in their 304 covering with connective tissue and encapsulation. In available literature only the internalization 305 in the intestine, but not excretion of particles larger than 1 µm has been described in vertebrates    For non-biodegradable microcapsules it is important to evaluate the long-term effects of their 342 introduction into various organs of fish. For this purpose, samples collected 1 day and 2-3 weeks 343 after injection were analyzed. The immune response fully develops by the day after the traumatic 344 effect (Antonio et al., 2015), so at this time it is possible to assess the general state of the fish 345 immune system. Between the second and third weeks after the injury, the acute phase of 346 inflammation ends, associated directly with the removal of damaged tissue structures, and the 347 regeneration process begins (Wahli et al., 2003). Therefore, if the immune response progresses 348 during this period, this indicates that immunogenic material is present in the body. 349 On the first day after the administration of PMs-PEG, the vast majority of microcapsules in D. 350 rerio sections were found spilled in the body cavity of the fish. Most microcapsules laid freely 351 between internal organs, but some of them were agglomerated inside large phagocytosing cells 352 with pale eosinophilic cytoplasm and a large basophilic nucleus. The phagocytosis of PMs-PEG 353 was observed outside the injection site and associated blood clot; thus, it did not result from non-354 specific phagocytosis of cell debris or infectious agents in the wound. This indicates that despite 355 the polyethylene glycol coating, PMs-PEGs were recognized by immune cells shortly after 356 injection. The obtained results are consistent with in vitro tests, which suggest that mammalian 357 phagocytes internalize a portion of microparticles coated with polyethylene glycol-containing PeerJ reviewing PDF | (2020:11:55227:1:1:NEW 23 Mar 2021) Manuscript to be reviewed 358 polymer within 4 hours (Wattendorf et al., 2008). Two weeks later all microcapsules in the body 359 cavity were engulfed by phagocytes, which became "foamed" because of a large number of 360 microparticles (Figs. 4A, 4B).

370
The recognition of the implant by the immune system leads to chronic inflammation in the 371 wound, which lasts until the foreign body is destroyed or isolated (Anderson et al., 2008). To 372 investigate the progression of the immune response provoked by PMs-PEG, we examined the 373 tissue microenvironment of the microcapsules stuck in different fish organs. In one of the fish 374 injected with PMs-PEGs, an abundance of microcapsules spilled into the body cavity and caused 375 considerable inflammation to the surface of the spleen and liver on the 14th day after injection 376 (Figs. 4C, 4D). The inflamed spleen increased in size and fused with the adjacent edge of the liver. 377 There were many microcapsules at the place of their confluence causing further inflammation. In 378 the liver close to the sites of PMs-PEG agglomeration, the parenchyma became loose due to its 379 replacement with phagocytic cells and eosinophilic granular cells (mast cells) (Fig. 4D), indicating 380 severe parenchyma dysfunction at the inflamed organ edge. Similar immune reactions were 381 observed in the kidneys two and three weeks after injection around the implanted fluorescent 382 material (Fig. 5C). There were many leucocytic cells surrounding the fluorescent microparticles; 383 however, given the hematopoietic function of the fish kidney, it has not been established whether 384 these cells represent normal renal parenchyma or were recruited due to inflammation. Both in the 385 liver and in the kidney, in the rest of each organ (i.e. free of microcapsules) hepatic and renal 386 parenchyma looked intact, with no signs of degeneration. This indicates that the irritant effect of 387 PMs-PEG occurred locally.

388
Dense nodules with fluorescent material consisting of several layers of phagocytic cells and 389 fibroblasts were found two weeks after PMs-PEG administration in the kidney and abdominal 390 cavity of the zebrafish (Figs. 5A, 5B, 5D). These specific morphological structures were 391 granulomas, formed as a result of attempts of immune cells to isolate the foreign bodies from 392 normal tissues (Anderson et al., 2008). The observed macrophage fusion around the implant with 393 its subsequent coating with a fibrin capsule is the final stage of the immune system response to the 394 irritant and occurs if the foreign body cannot be split and removed from the tissue. Therefore, the 395 immune response to PMs-PEG delivered to fish organs through the circulatory system was shown 396 to develop according to the classical foreign body reaction scenario. The same outcome of the 397 healing process was observed previously in zebrafish muscles 22 days after intramuscular injection On the fish gill sections, the microcapsules were found only as separate fluorescent objects 413 in blood capillaries, never as agglomerates. Unfortunately, it was difficult to determine whether 414 they were engulfed by phagocytes, as a result of the destruction of blood cells by Buen's fixative. 415 However, no signs of local inflammation were observed around them in gill lamelae.

417
Comparative morphology of fish kidney 418 As shown above, PMs-PEGs are immunogenic, therefore they can potentially cause not only 419 local inflammation, but overstimulation of the immune system with severe toxic effects on the 420 whole organism. In a hyperinflammatory state, tissue damage, functional organ failure and, 421 ultimately, death can occur as a result of an uncontrolled cascade of inflammatory reactions (Philip 422 et al., 2017). To assess the potential toxicity of microcapsules in vivo, we compared the 423 morphology of organs of D. rerio that received an injection of either PMs-PEG or saline 424 (Supplementary Table 1). 425 The D. rerio kidney is an abdominal organ located along the spine of zebrafish and consists 426 of tissues with mixed hematopoietic, immunologic, endocrine, and urinary functions (Wolf et al., 427 2015). The renal tissue of examined fish was represented by parenchyma from undifferentiated 428 cells with condensed basophilic nuclei. Throughout the entire length of the parenchyma, there were 429 numerous distal and proximal renal tubules and filtering renal glomeruli (renal corpuscles) (Fig.  430 6A).

431
Injection of PMs-PEG or saline was performed into the central bulk of the fish trunk kidney. 432 At the injection site, the tissues were mechanically damaged, and over the next three weeks, 433 successive stages of the healing process were observed there. Where the renal artery was ruptured, 434 blood spilled into the body cavity and formed a blood clot, which was gradually disassembled by 435 relatively rare phagocytic cells that absorb fragments of blood cells. The hematoma in the 436 abdominal cavity remained not resolved completely for three weeks after the injection.

Manuscript to be reviewed
Such pathologies as single nephron degeneration and vacuolar degeneration (Fig. 6D) of the 478 distal tubule were sporadically found in fish from the control group. The clots of organic material 479 were also periodically detected in the lumen of the renal tubules of fish that received saline. At the 480 injection site, the lumen of the tubule was filled with granular material, probably fragments of 481 destroyed cells (Fig. 6C). In one individual from the control group, a neoplastic process associated 482 with massive lesion of renal tissue was found in the head kidney on day 14 after injection (Fig.  483 6F). In a rather large field, the renal parenchyma was replaced by spindle anaplastic cells with 484 eosinophilic cytoplasm and large basophilic nuclei. Degeneration and extensive necrosis of the 485 renal tubules and glomeruli, together with massive involvement of phagocytes and eosinophilic 486 granule cells, indicated ongoing acute inflammation in the affected area.

487
In the kidney of fish that received PMs-PEG, the injection site was a lot of shapeless loose 488 material, consisting of coagulated red blood cells mixed with microcapsules. Two and three weeks 489 after injection, severe inflammation and foreign body reaction were observed around the 490 fluorescent material, as described above (Figs. 5C, 5D). In the regions remote from the injection 491 site the normal architecture of trunk kidney was generally maintained: the distal and proximal renal 492 tubules and renal glomeruli were clearly distinguishable; the parenchyma between them was 493 represented by hematopoietic tissue, consisting of immature lymphocytes, as well as leukocyte 494 precursors and erythrocytes (Fig. 7A). Fish from the control group and fish treated with PMs-PEG statistically did not differ in the 505 level of estimated morphological parameters. On the second week after injection, in the renal 506 parenchyma and around the glomeruli of fish that received microcapsules, the number of 507 pigmented macrophages was higher than in fish of the control group (Fig. 5C), which indicates an 508 increase in the inflammatory process. Macrophages directly absorbed microcapsules and were 509 involved in the absorption of damaged erythrocytes or nephrons. However, on the third week after 510 injection, the opposite effect was observed: the number of macrophages found in the renal 511 parenchyma was even lower in fish from the experimental group than in fish from the control 512 group. This probably indicates that three weeks after the injection, the generalized inflammation 513 ceased, while the foci of inflammation remained directly around the aggregates of the 514 microcapsules.

515
On the second and third week after capsule administration foci of early stages of renal 516 glomerular fibrosis (Fig. 7B) and dilated capillaries in the renal glomerulus (Fig. 7A) were PeerJ reviewing PDF | (2020:11:55227:1:1:NEW 23 Mar 2021) 517 observed in the wound of the site of injection. Foci of intense necrosis of individual renal tubules 518 were also found in fish from the experimental group on the 22 nd day after injection (Figs. 7C, 7D). 519 The severity of these conditions can be considered mild since they were sporadic and did not affect 520 the adjacent renal tissues, which retained their ability to function normally.

521
Such morphological pathologies such as blood clots, vascular fibrosis, renal glomerular 522 necrosis, and fibrosis, as well as granuloma formation, were not found outside of the injection site. 523 The number of new immature nephrons, a diagnostic indicator of the regeneration process, did not 524 differ significantly in the fish kidneys of either the control or the experimental group. Therefore, 525 the inflammatory process in the response of D. rerio kidney to the PMs-PEG administration can 526 be characterized as moderate and local.

528
Comparative morphology of fish gills 529 Normal morphology of gills was observed in the fish from both control and experimental 530 groups (Fig. 8A). The organ consisted of cartilaginous gill arches with outgrowths of gill filaments 531 with two rows of secondary lamellae. Between lamellae, filaments were coated with 1-3 rows of 532 squamous pavement cells. No disturbances in the structure of filaments, such as massive 533 hyperplasia of filament epithelium, fusion of filaments, fibrinolysis, and leukocyte infiltration, 534 were found. There were also no vasculature disorders, with the exception of several cases of 535 hyperemia of gill filaments. Epithelial folds that were secondary lamellae were well separated from 536 each other and in cross-section looked like a thin channel formed from 1-2 layers of squamous 537 cells and filled with red blood cells lined up in a row. In fish from the control group some small 538 zones of lamellar epithelial hyperplasia were noted. On the 14th day after injection, abundant 539 extracellular material rich in eosinophilic protein, presumably mucus (Saleh et al., 2018; Sveen et 540 al., 2019), was found on the surface of the lamellae in two reference fish (Fig. 8B). This gill 541 covering is likely to perform a protective function, reducing the availability of the gill epithelium 542 to irritating agents from the external environment, and its secretion by epithelial cells may be an 543 artifact associated with the use of clove oil as an anesthetic. However, no severe lesions of the 544 lamella structure as lifting or desquamation of the respiratory epithelium, aneurysms, and other 545 circulatory disorders were found in fish that received an injection of saline.  A comparative analysis of the gill microstructure after saline or PMs-PEG introduction into 555 the kidney did not reveal significant differences in the frequency or area of the studied 556 histopathologies. This indicates a relatively small effect of systemically administered PeerJ reviewing PDF | (2020:11:55227:1:1:NEW 23 Mar 2021) 557 microcapsules on the structure and function of this organ. A rather mild violation of the structure 558 of gill tissue, such as the fusion of individual secondary lamellae, was detected in all fish three 559 weeks after injection of PMs-PEG. This was either direct adhesion of neighboring lamellae or 560 filling of interlamellar sulci due to epithelial hyperplasia. As hyperplasia lasted only for a 561 couple/several neighboring lamellae, without forming continuous fusion, therefore the lesion area 562 was too small to seriously decrease in the surface of gas exchange and impair respiration. Lamella 563 fusion is often a non-specific response to chronic inflammation that causes proliferation of a mixed 564 population of the pavement, mucous and chloride cells and/or leukocyte infiltration (Wolf et al., 565 2015). In PMs-PEG-injected fish, sporadic respiratory epithelial hyperplasia may be caused by a 566 general increase in immunoreactivity due to the irritating effect of the microcapsules.

567
Two and three weeks after PMs-PEG injection, a slight increase in the frequency of lifting 568 and desquamation of the epithelium (Supplementary Table 1) that line the blood capillaries was 569 recorded. Epithelial lifting is the initial stage preceding the rupture and loss of the epithelium, 570 which is considered as a severe pathology since it directly violates the respiratory process. In fish 571 injected with microcapsules, such changes were sporadic and could not significantly affect the 572 respiration of the fish, especially considering regenerative capabilities of gill tissue (Wolf et al., 573 2015). Also, many authors report that such violations are often artifacts, a consequence of 574 mechanical damage to tissue during the cutting of histological sections (Frasca et al., 2018;Wolf 575 et al., 2015). 576 Since the gills are intensively supplied with blood, one of the expected adverse effects from 577 the systemic administration of a large number of PMs-PEG was the agglomeration of 578 microcapsules in the finest gill capillaries. Formation of blood congestions followed by severe 579 anoxia can cause direct toxicity of the microimplants for the fish. Within three weeks after 580 microcapsule administration, we revealed no signs of circulation obstructions (aneurysms, blood 581 clots, etc.) in primary and secondary gill filaments. The elastic polymer shell of the PMs, which 582 allows them to squeeze through narrow spaces, probably contributes to their satisfactory 583 hydrodynamic characteristics in fish capillaries.

585
Comparative morphology of fish liver 586 The hepatic structure of fish from the control group, which received an injection of saline 587 into the kidney, was typical for zebrafish (Cheng, 2004). We observed homogeneous parenchyma 588 consisted of dense, uniform, angularly shaped hepatocytes with soft margins with a large 589 basophilic nucleus in the center and a very vacuolated, clear cytoplasm (Fig. 9A). Large vacuoles 590 accumulating glycogen have been reported to be typical for zebrafish hepatocytes, especially for 591 captive fish fed with commercial feed (Wolf, Wheeler, 2018). There occasionally were small 592 basophilic stained macrophages (Kupffer cells) squeezing between hepatocytes (Figs. 9A, 9C). 593 Within the hepatic parenchyma, there were many sinuses and blood vessels moderately filled with 594 red blood cells and bile canaliculi lined with basophilic epithelial cells. The presence of 595 hepatocytes with pyknotic nuclei was detected in many fishes from the control and experimental   Figure 6 Morphological structure of zebrafish kidneys after injection of saline into fish kidney.