Proceedings of the 2023 Division of Translational Toxicology Satellite Symposium

Dr. Gregory A. Krane (Moderna, Cambridge, Massachusetts) presented three studies with the goal of suggesting appropriate terminology for accumulations of round cells. All of the studies were for development of lipid nanoparticle-encapsulated (LNP), messenger RNA (mRNA) therapeutics, although the mRNA product and specific indication for each study were de-identified for the purposes of the symposium. For each study, Dr. Krane provided relevant background information, presented a series of photomicrographs, and then queried the audience with a series of voting choices, followed by explanation of the diagnosis chosen by Moderna’s pathology team. Dr. Krane would like to acknowledge Franny Ambrose and Dr. Annie Golding of Moderna, who performed the immunohistochemistry work described in this manuscript, and Dr. Youmna Kfoury of Moderna, who provided the gross photographs of the spleen for study three ( Figures 4S - T ). Dr. Krane also thanks Beth Mahler of Experimental Pathology Laboratories, Inc. for support with figures, Dr. Martha Hensel of the MD Anderson Cancer Center for her peer review of the manuscript, and Dr. Erin Quist of Charles River Laboratories, Inc. and the pathology team at Moderna (especially his manager Dr. Eric Jacquinet) for their feedback with regards to the development of this presentation and manuscript. Study one featured two dose range finding studies for a cytokine intended for oncology indications, with one study being performed in 2.5 – 3 year old male and female cynomolgus monkeys (NHP), and the other study being performed in 9 – 10 week old male and female Sprague Dawley rats. Animals were administered the test article once weekly for three weeks followed by terminal necropsy 24 hours after the final administration. Dr. Krane showed photomicrographs of histologic sections from NHP liver ( Figures 4A - B ), rat liver ( Figures 4C - D ), rat adrenal gland ( Figures 4E - F ), and rat and NHP livers from a prior study with a similar test article ( Figures 4G - H ). All images shared a common finding of a monomorphic population of densely packed, medium sized round cells with a prominent nucleus with open chromatin and small amounts of cytoplasm. This expansile population of cells compressed and sometimes infiltrated and replaced the surrounding parenchyma. In the liver, although predominantly in portal and peri-portal areas, these cells were also randomly present amongst hepatocytes and infiltrated sinusoids, were increased in number in the vasculature, and were associated with single cell necrosis of hepatocytes. In the rat, this finding was diagnosed in the liver, adrenal gland, pancreas, lung, epididymis, and Harderian Gland, and in the NHP in the liver, adrenal gland, pancreas, thyroid gland, uterus, and choroid plexus of the brain (although images from each affected organ were not presented to the audience, the audience was shown a list of affected organs). Relevant findings shared with the audience (without presented photomicrographs) were increased myeloid cellularity in the bone marrow of the rat, a shift of myeloid precursors toward early stages of proliferative cells (bone marrow cytology) in the rat, increased macrophage cellularity in the red pulp and increased lymphocyte cellularity in the white pulp of the rat spleen, and increased macrophage cellularity of lymph nodes in the rat and NHP. Voting choices for Study one, with voting results were as follows: infiltration, lymphocytic (6%), increased cellularity, lymphocytic (1%), infiltration, mononuclear cells (59%), increased cellularity, mononuclear cells (13%), inflammation, mononuclear cells (4%), infiltration, mixed cells (9%), inflammation, mixed cells (0%), lymphoma (0%), and infiltration, myeloid (9%). The diagnosis chosen by Moderna’s pathology team was infiltration, myeloid. The Moderna pathology team found themselves in a difficult situation for this study, as there was not an INHAND or SEND diagnosis compatible with the features of this finding. This monomorphic population of densely packed round cells exhibited a leukocyte morphology, but it was not consistent with any particular mature leukocyte lineage. The morphology and biologically aggressive behavior differed from mononuclear or mixed cell infiltration as described in INHAND guidelines. The finding was not inflammatory in nature, as there were no tissue responses consistent with inflammation, nor did it efface tissue architecture to the point of being able to justify it as a neoplasm. The team felt this was an infiltration (versus a diagnosis of increased cellularity), as the cells were in aggregates and not throughout an entire tissue compartment. The team was concerned that use of the term “infiltration, mononuclear” would not capture the aggressive behavior of the cells infiltrating, compressing, and remodeling architecture of the host tissue parenchyma. Given the resemblance of the cells to myeloid precursors in bone marrow, combined with the increased cellularity of cells with myeloid lineage in some hematolymphoid tissues and the presence of histiocytic cells in this infiltrate in a prior study, the Moderna pathology team carried forward the diagnosis of “infiltration, myeloid” that had been utilized in the prior study. Given the team’s concern for neoplastic potential of this myeloid infiltration exhibiting biologically aggressive behavior, the investigative pathology group followed with immunohistochemistry to characterize the cell type and better evaluate risk. Dr. Krane presented immunohistochemistry images from NHP liver for CD3, CD8, NKG2α, and CD20, which immunolabel T cells, cytotoxic T cells, NK cells, and B cells, respectively, and rat liver for CD3 and CD8. These images demonstrated the cellular infiltrate to predominantly immunolabel for CD3 and CD8, which characterized them as a population of cytotoxic T cells. In the NHP, there were fewer cells immunolabeled for NKG2α (NK cells) and CD20 (B cells). Dr. Krane presented quantitative image analysis data that reinforced the interpretations of the immunohistochemistry. Given the characterization of these infiltrating cells as predominantly cytotoxic T lymphocytes, the finding was amended to “infiltration, lymphoid,” as lymphocytes are not derived from the myeloid lineage 25 . Although the interpretation of the cells’ behavior was unchanged, their characterization was useful in interpreting the finding to be a result of exaggerated pharmacology, with stimulation of both lymphocytes in peripheral tissues and myeloid cells in bone marrow and hematolymphoid tissues in the same mechanism as the intended pharmacology, but to an amplified extent as compared to the desired clinical effect. This was influential in the decision to use lower doses for follow-up GLP studies. In summarizing this study, Dr. Krane mentioned that it is important to use INHAND and SEND terms whenever possible, but this example represents a unique circumstance where pathologists must use their scientific judgement to create a new term (while justifying the terminology in their narrative) that best captures the features of the finding. Study two showed findings from a 9 – 10 week old female Sprague Dawley rat in a dose range finding study with an unidentified test article. This rat was found dead on study with no prior clinical observations, and the only gross finding at necropsy was bilaterally enlarged iliac lymph nodes. Dr. Krane showed photomicrographs of histologic sections of the thymus ( Figures 4I - J ), lymph node ( Figures 4K - L ), and ovary ( Figures 4M - N ), with corresponding images from a control animal on the same study for comparison. In all of the images shown, there sheets of monomorphic, small to medium round cells with hyperchromatic nuclei with small amounts of cytoplasm that variably effaced and replaced pre-existing architecture of the tissues. Large aggregates of similarly appearing lymphocytes were also present in the vasculature of various tissues. Other organs affected by this finding, but not shown in the presentation, were the lung, kidney, uterus, cervix, and vagina. Voting choices for Study two, with voting results were as follows: infiltration, lymphocytic (14%), increased cellularity, lymphocytic (24%), infiltration, mononuclear cells (20%), increased cellularity, mononuclear cells (16%), inflammation, mononuclear cells (0%), infiltration, mixed cells (0%), inflammation, mixed cells (2%), and lymphoma (24%). Due to the multi-organ effacement of pre-existing architecture by a monomorphic round cell population, the correct diagnosis was lymphoma. After votes were cast, Dr. Krane showed photomicrographs demonstrating immunohistochemistry for CD3 (diffuse immunolabeling of neoplastic cells, T cell marker, Figure 4O ) and CD20 (no immunolabeling of neoplastic cells, B cell marker, Figure 4P ), further justifying the diagnosis of lymphoma plus additionally characterizing it as a malignancy of T cells. Study three showed findings from 10 week old, female C57Bl6 mice that were intravenously administered a test article intended to be part of a bone marrow transplantation preparation regimen three times during one week immediately followed by terminal necropsy. Relevant findings were increased liver enzymes, grossly enlarged spleens at necropsy, and increased spleen weights. Only the liver and spleen were collected at necropsy for histologic examination by the principal investigator. Histopathology from the liver and spleen shown in the presentation demonstrated in both tissues a population of monomorphic, medium to large round cells with prominent nuclei and variable amounts of cytoplasm, which sometimes were multinucleated or exhibited mitoses. In the liver ( Figures 4Q - R ), this population of cells filled and expanded hepatic sinusoids and compressed the adjacent parenchyma, with occasional concurrent hepatocellular degeneration/necrosis in these neighboring regions. In the spleen ( Figures 4S - V ), this population of cells filled the red pulp and obscured but did not efface background architecture. This population of cells correlated with the grossly enlarged spleens and increased spleen weights. Voting choices for study three, with voting results were as follows: infiltration, lymphocytic (4%), increased cellularity, lymphocytic (9%), infiltration, mononuclear cells (13%), increased cellularity, mononuclear cells (7%), inflammation, mononuclear cells (7%), infiltration, mixed cells (4%), inflammation, mixed cells (2%), and lymphoma (55%). The diagnosis chosen by the Moderna pathology team for study three was increased cellularity, mononuclear. Dr. Krane acknowledged the majority of the audience favored the diagnosis of lymphoma, though the Moderna pathology team did not feel there was enough remodeling of tissue architecture by the cellular population to justify a diagnosis of neoplasia. However, the principal investigator of this study was cautioned that this diagnosis was made solely on the evaluation of the two tissues submitted. Should other tissues from the animals have been collected and made available for microscopic examination, depending on the extent of change in tissue architecture in other tissues, it is possible that the diagnosis for this study may have been lymphoma instead of increased cellularity, mononuclear. Dr. Krane discussed that evaluation of limited tissue sets is not an uncommon situation for pathologists, especially for discovery studies where they become involved after most of the project has already been completed. He discussed the role of pathologists in team science and encouraged attendees to educate principal investigators to involve pathologists early on in study design meetings to avoid scenarios where additional information could be helpful to optimize study conclusions and recommendations but is not possible to ascertain due to deficiencies in study design such as inadequate sample collection for clinical pathology, histopathology, organ weights, or other endpoints. An audience member personally discussed with Dr. Krane that perhaps for these scenarios, it is more appropriate that a pathologist use the terminology of “impression” or “opinion” instead of “diagnosis,” as is the practice of some health care providers in diagnostic radiology or pathology settings, given that such language characterizes the limitations of the interpretation based on the available information. The theme of this presentation was terminology of findings characterized by accumulations of round cells, with a focus on infiltration, increased cellularity, and round cell malignancy. Dr. Krane provided a summary table ( Table 5 ) comparing features of these findings, based on a combination of anecdotal experience and available literature 26 . Important points to consider when considering these diagnoses are distribution of the cells, percentage of the tissue affected, maintenance of tissue architecture, if the cells have heterogenous or monomorphic appearance, if the cells are in abnormal anatomic locations, the number of tissues affected, if concurrent inflammation is present, and the mitotic rate. Although one audience member postulated that infiltration and increased cellularity may be the same finding and should have a unified term, Dr. Krane did feel that they appear distinct enough visually to be diagnosed as separate findings, even if the underlying mechanism causing the finding is the same. Representatives from the INHAND oversight committee agreed with Dr. Krane’s response in personal discussions. Primarily, distinctions between infiltration and increased cellularity are that infiltration tends to represent multifocal aggregates of cells, whereas increased cellularity tends to represent cells occupying an entire tissue compartment (and sometimes expanding that compartment). Although the percentage of tissue affected may be greater with increased cellularity as compared to infiltration, otherwise the criteria for these findings are quite similar. 26 However, round cell malignancies differ from both infiltration and increased cellularity, as they tend to affect larger amounts of tissue, efface and/or replace pre-existing tissue architecture, are composed of monomorphic cell populations, occupy abnormal anatomic location, affect greater numbers of organs, may have secondary inflammation, and exhibit a higher mitotic rate. 26 Ultimately, the primary take home message of Dr. Krane’s presentation is that the diagnoses of infiltration, increased cellularity, round cell tumor, or inflammation must be carefully weighted, as they represent different histopathologic presentations, even if they may share some underlying mechanisms. Although utilization of SEND and INHAND compatible terminology is recommended whenever possible, sometimes it is necessary to create alternative scientifically justifiable terminology if there is not a SEND or INHAND finding that best conveys the appearance, mechanism, and/or implications of the finding. Identification of such safety findings can help drive partnerships with investigative studies that involve scientists in other disciplines that provide further context to such findings, and it is imperative that pathologists build relationships with such scientists to design studies that can provide the most impactful and useful data possible. Ultimately, training both as veterinarians and pathologists provides comparative biomedical science knowledge from the subcellular to the multi-animal level that can create opportunities to improve the value delivered by scientific collaborations with a variety of stakeholders who depend on us to provide safe, effective, and novel diagnostics and therapies to patients with unmet need.

Dr. Shawn Lennix (formerly of Labcorp Early Development Laboratories Inc., Madison, Wisconsin) presented a case featuring an oral mass in a rat from a 2-year toxicity/carcinogenesis study for which Dr. Molly Boyle (Somerset, New Jersey) was the study pathologist. Dr. Lennix thanked Drs. Victoria Laast, Melissa Behr, and Jennifer Dreyfus (Madison, WI) for their support; Steve Van Adestine and Jasmine Holland for providing photography and digital imaging assistance; and Susan Lynk and Nate Sendelbach for performing immunohistochemistry. The case presentation featured a 20-week-old male, Wistar Han Rat (RccHan ® :WIST) control animal receiving deionized water on a 2-year toxicity/carcinogenesis study. On Day 92/Week 14 of the study, a 29 x 20 x 20 mm soft, dark red mass was noted on the mandible and the animal was euthanized for welfare reasons. At necropsy, the mandibular lymph nodes were grossly unremarkable. Upon histopathologic examination, the axillary and mesenteric lymph nodes had no microscopic findings. Dr. Lennix showed the audience images of the mass in situ ( Figures 6A and 6B ), low and high magnification photomicrographs of representative regions of the mass stained with H&E ( Figures 6C though 6J ), and tissue immunolabelled with pancytokeratin and vimentin ( Figures 6K though 6L ). The audience was asked to diagnose the mass and the voting choices and results were as follows: ameloblastoma (39%), ameloblastic odontoma (32%), ameloblastic carcinoma (28%), odontoma (1%), squamous cell carcinoma (0%), and melanoma (0%). Dr. Lennix was visibly pleased some participants agreed with her diagnosis of ameloblastic carcinoma and a variety of other answers were selected by participants. Dr. Lennix walked through features of the tumor in this case. She pointed out odontogenic epithelial cells arranged in islands and trabeculae that surrounded a large cystic space, replaced bone, and compressed adjacent skeletal muscle ( Figures 6C – 6D ). The neoplastic cells were arranged in highly cellular aggregates with irregular margins supported by variably dense fibrovascular stroma ( Figure 6E ). At the periphery of the islands, cuboidal to columnar neoplastic odontogenic epithelial cells with nuclei occasionally in anti-basilar locations were arranged in undulating or bosselated margins forming “ink drop” patterns ( Figure 6E ). At the center of islands, neoplastic cells were spindle-shaped and formed streams and bundles (“sarcoma-like”) or were polygonal and reminiscent of stellate reticulum with variably distinct fine intercellular processes and polygonal nuclei ( Figures 6G – 6H ); in some regions, the neoplasm exhibited cystic degeneration ( Figure 6I ). Multiple cysts lined by parakeratinized stratified squamous epithelium contained free erythrocytes, hemosiderophages, degenerate epithelial cells, and neutrophils. In a few areas, cells at the center of islands showed squamous differentiation and some small aggregates of were keratinized ( Figure 6I ). There were18 mitoses per 2.37 mm 2 and, interestingly, mitotic figures were observed both in basilar cells and those cells more centrally located within islands. Cells demonstrated mild anisocytosis and anisokaryosis ( Figure 6J ). Vascular and neural invasion were not observed. The neoplastic cells had positive immunoreactivity to both cytokeratin and vimentin with moderate and strong cytoplasmic staining, respectively ( Figures 6K – 6L ). Next, Dr. Lennix discussed the diagnostic features of proliferative lesions associated with teeth and referenced a number of sources 36 – 46 as summarized in Table 7 . She noted the specific features important in diagnosis of the tumor which include 1) identification of odontogenic epithelium, 2) stellate reticulum, 3) a determination on the presence of dental hard tissues, and 4) mitotic figures. Typically, odontogenic epithelial cells demonstrate ameloblastic morphology characterized by palisading cuboidal to columnar epithelial cells with anti-basilar nuclei (i.e., reversed polarity) and clear space in the basilar cytoplasm. Stellate reticulum is composed of polygonal-shaped cells separated by faintly staining matrix or clear space with delicate cell-to-cell cytoplasmic processes. One or both of these tissue arrangements should be present in odontogenic tumors and are variably distinct. In this particular neoplasm, the odontogenic epithelium was loosely ameloblastic with many areas of jumbled epithelial cells with nuclei that failed to regularly demonstrate the characteristic palisading of ameloblastoma. A well-defined stellate-like reticulum was only present in a few regions. Dental hard tissues were not seen in the neoplasm and thus ruled-out odontoma and ameloblastic odontoma. Regarding odontogenic tumors, Dr. Lennix mentioned her experience in diagnosing canine ameloblastoma has included the observation of low mitotic counts that are frequently 0–1 mitotic figures in 10 high power (400x) fields. As a result, the more robust mitotic count in this case raised her suspicion that a diagnosis other than ameloblastoma should be considered. Dr. Lennix briefly mentioned the presence of squamous differentiation and keratinization are not features specific to oral squamous cell carcinoma and have been observed in ameloblastic carcinoma. 47 Dr. Lennix briefly discussed the diagnostic criteria for dental tumors in INHAND 38 and compared the features of odontoma, ameloblastic odontoma, and ameloblastoma to ameloblastic carcinoma. Dr. Lennix reiterated that odontogenic epithelium and stellate reticulum, while reliably distinct in ameloblastoma, were not easily identifiable in this case, thus adding greater suspicion to her hunch that another diagnosis should be considered. Dr. Lennix then gave an overview of publications related to select odontogenic tumors in the veterinary and toxicologic pathology literature ( Table 8 ). Odontogenic tumors have been induced in mice by polyoma virus. 40 Experimentally-induced odontogenic carcinoma has been reported in Wistar rats. 44 Spontaneous ameloblastoma has been described in rats and has been considered rare 38 , 43 and, similarly, spontaneous ameloblastic carcinoma has been reported in a dog, 42 a horse, 37 and a domestic brown rat. 47 In Fischer rats, ameloblastic carcinoma has been induced by methlynitrosourea. 36 In dogs, a diagnosis of ameloblastic carcinoma relies on microscopic features such as increased mitotic rate, loss of odontogenic features, cellular pleomorphism, and biologic behavior (i.e., metastatic disease). 45 , 46 Descriptions of ameloblastic carcinoma have included the features of high nuclear to cytoplasmic ratio, dark eosinophilic or vacuolated cytoplasm, nuclear pleomorphism with large to giant and pyknotic or small, hyperchromatic nuclei. 37 , 42 Commentary within the veterinary literature is sparse regarding mitotic activity with mitoses of 6–9/hpf and 5–6/hpf reported in ameloblastic carcinoma in a dog and horse, respectively. 37 , 42 Dr. Lennix then concluded with a summary of the case and take home points. The histopathologic distinction between ameloblastoma and ameloblastic carcinoma can be a difficult one to make. Despite having some features relatable to ameloblastoma, a diagnosis of ameloblastic carcinoma was favored based on the presence of odontogenic epithelium, stellate-like reticulum, cellular pleomorphism and increased mitotic figures. Cellular pleomorphism was characterized by a combination of neoplastic cells exhibiting polygonal and spindle-shaped morphology in a pattern of streaming bundles that gave it a “sarcomatous” appearance. The co-expression of vimentin and cytokeratin has been reported in ameloblastic carcinoma in a domestic brown rat47 and has been described in human ameloblastoma and fetal tooth germ. 6 Additionally, in humans, ameloblastic carcinoma reliably expresses CK18, MMP-2, and MMP-9 and have an increased Ki67 index which distinguishes it from ameloblastoma. 48

Dr. Allison Boone (Experimental Pathology Laboratories (EPL), Inc., Durham, North Carolina) presented 3 cases of liver lesions observed in B6C3F1/N mice. Cases were derived from two, 2-year toxicology/carcinogenesis bioassays conducted by the Division of Translational Toxicology/National Institute of Environmental Health Sciences (DTT/NIEHS). The study pathologists for the bioassays were Drs. Abraham Nyska and the late J.C. Peckham. Dr. Boone would like to thank Drs. Arun Pandiri and Mark Cesta at NTP/NIEHS, as well as Michael Carden, Leslie Couch, Emily Singletary, and Lorri Ezedin at EPL, Inc. for their assistance and support with the presentation. Attendees were shown a series of both low and high-magnification photomicrographs of the liver lesions for all 3 cases ( Figure 3 ). The diagnostic choices were similar for all cases and Table 3 includes the voting choices and results. Case 1 comprised photomicrographs from a focus of cellular alteration ( Figure 3A - D ). The example consisted of a well-demarcated basophilic focus with hepatic cords that merged imperceptibly with adjacent hepatocytes ( Figures 3A and 3B ). Additionally, the focus exhibited minimal cellular atypia ( Figure 3C ) and low mitotic activity was observed ( Figure 3D ); compression/invasion of the hepatic parenchyma, necrosis, and hemorrhage was not observed. Only 25% of attendees selected the diagnosis of focus of cellular alteration, and 29% and 39% of attendees chose hepatocellular nodular hyperplasia and hepatocellular adenoma, respectively, as their voting choices. These results demonstrate the challenge in differentiating neoplastic and nonneoplastic hepatic lesions. For cases 2 and 3, Dr. Boone presented photomicrographs from a hepatocellular adenoma ( Figure 3E - H ) and hepatocellular carcinoma ( Figure 3I - L ), respectively. The hepatocellular adenoma consisted of a well-demarcated mass ( Figure 3E ) that compressed adjacent hepatic cords ( Figure 3F ). The mass lacked retention of lobular architecture ( Figure 3G ) and mild to moderate cellular atypia was observed in the neoplastic population ( Figure 3H ). The hepatocellular carcinoma replaced greater than 50% of the liver section ( Figure 3I ), compressed adjacent hepatic cords, and contained a small area of necrosis ( Figure 3J ). Neoplastic hepatic plates were up to 6 cells thick, lacked normal lobular architecture ( Figure 3K ), and exhibited marked cellular atypia ( Figure 3L ). The most prevalent voting results for cases 2 and 3 agreed with the NTP diagnoses of hepatocellular adenoma and carcinoma, respectively. Following voting, Dr. Boone reviewed the features specific for each diagnosis with a collage of the photomicrographs as seen in Figure 3 and a summary of the diagnostic features specific to each lesion based on the International Harmonization of Nomenclature and Diagnostic Criteria (INHAND) guidance document ( Table 4 ) 16 . During the beginning of the discussion, Dr. Boone reminded attendees that foci of alteration and hepatocellular tumors can occur spontaneously or secondary to chemical exposure with differential species/strain susceptibility. For the remaining component of the discussion, Dr. Boone focused on the significance of the foci of cellular alteration in rats and mice. She mentioned how all foci may not be viewed as a preneoplastic change 17 – 20 ; however, there are reports that have associated the foci as precursors to neoplasia 21 , 22 . Given this dilemma, Dr. Boone mentioned ways to further characterize the relevance of the foci in rodent studies for hazard assessment, which included reversibility of the lesion after cessation of exposure during a recovery period, consideration of enzyme markers of preneoplasia (e.g., gamma-glutamyl transpeptidase (GGT)) 23 , and the mode of action of the associated toxicant. Furthermore, she mentioned newer technological advances that may enable further characterization of the foci including spatial omics, single cell sequencing, and error corrected duplex sequencing. Dr. Boone concluded the presentation with an update on the current activities in the Division of Translational Toxicology, Comparative and Molecular Pathogenesis Branch, Molecular Pathology Group (MPG), which may assist with further characterization of the foci. She mentioned that the MPG is currently embarking on a project that involves whole genome sequencing of rodent tumors/non-tumors resulting from various exposures 24 and generating a database of mutation signatures and cancer driver genes. The studies will not only assist in the development of targeted sequencing panels to investigate preneoplasia but will also aid in the comparison of human and rodent cancers induced by similar exposures, and thus establish translational relevance with human cancers.

Dr. Erin M. Quist (Charles River Laboratories, Inc., [CRL] Durham, North Carolina) presented several interesting cases recently reviewed by a Pathology Working Group (PWG) as part of a special review conducted by the Division of Translational Toxicology, National Institute of Environmental Health Sciences (DTT/NIEHS). The presented cases featured neoplastic lesions from B6C3F1/N mice that were part of a 2-year toxicity/carcinogenicity study in which Dr. Michael Ryan (Battelle, Columbus, Ohio) was the Study Pathologist (SP), the Quality Assessment Pathologist (QAP) was Dr. Crystal Johnson ([then] CRL, Durham, North Carolina), and the Reviewing Pathologist (RP) and PWG Coordinator (PWGC) for the special review was Erin Quist of Charles River Laboratories (Durham, North Carolina). Thanks to Dr. Charan Ganta and the NIEHS Histology Core Laboratory for assistance with immunohistochemistry (IHC), Emily Singletary and Leslie Couch for materials acquisition, whole slide imaging and PWG technical support, and Beth Mahler for photographic editing. For all 4 cases, a series of photomicrographs were presented to the audience that included both low- and high-power magnifications of select sections of liver from the 2-year study ( Figure 1 ). Dr. Quist informed the audience that the diagnostic choices would be the same for all 4 cases and that the audience would be asked to revote after viewing IHC images for each case; the voting choices and results are presented in Table 1 . Case 1 provided an example of liver – hemangiosarcoma ( Figures 1A – 1H ). The liver section was largely thrombosed with a poorly demarcated, unencapsulated neoplasm along the periphery that was characterized by poorly formed, haphazardly arranged, vascular channels lined by pleomorphic endothelial cells ( Figures 1A – 1D ). Vascular channels often expanded sinusoidal areas, disrupting the normal lobular architecture and creating large, blood-filled spaces in between the remaining hepatic cords ( Figures 1C – 1D ). Neoplastic cells occasionally formed blind-ended trabeculae that could be observed at a higher magnification ( Figure 1D , arrowhead). Mitotic figures were rare. To help differentiate the neoplastic cell type and confirm the diagnosis of liver – hemangiosarcoma, the following IHC stains were used: cluster of differentiation (CD)31 (endothelial cell marker) and cytokeratin (CK)18 (hepatocellular marker). Neoplastic cells exhibited positive CD31 staining ( Figures 1E – 1F ) but were negative for CK18 ( Figures 1G – 1H ), while remaining hepatocytes stained intensely positive with CK18 which served as a great internal, positive control ( Figure 1H ). Case 2 featured an example of liver – hepatoblastoma in another, largely thrombosed liver section ( Figures 1I – 1N ). The neoplasm was well-demarcated and confined to the periphery by a thick, fibrous connective tissue capsule that separated the neoplasm from the adjacent necrotic area ( Figures 1J – 1L ). Neoplastic cells formed what appeared to be an irregular rosette structure with a central, dilated, vascular channel surrounded by several cell layers of radially arranged, small, intensely basophilic, elongated cells ( Figures 1J – 1L ). Mitotic figures were frequently observed. On IHC, neoplastic cells were negative for both CD31 ( Figure 1M ) and CK18 ( Figure 1N ), the negative results being typical of hepatoblastoma, underscoring the major challenge of diagnosing hepatoblastoma with IHC. Cases 3 and 4, unbeknownst to the audience, were two different liver tumors from the same animal: a hepatocellular carcinoma ( Figures 1O – 1Q , 1U – 1X ) and hemangiosarcoma ( Figures 1O – 1T , 1Y – 1Z ) with metastasis to the hepatocellular carcinoma section ( Figures 1R and 1X ). The hemangiosarcoma presented for Case 3, featured yet another thrombosed liver section containing a poorly demarcated, unencapsulated, neoplasm within the periphery of the tissue that was composed of haphazardly arranged vascular channels ( Figures 1S – 1T ). Neoplastic cells were pleomorphic endothelial cells similar in appearance to those presented in Case 1, and stained positive with the endothelial cell marker, CD31 ( Figure 1Y ); neoplastic endothelial cells were negative for CK18, but as in Case 1, remaining hepatocytes were intensely positive for this hepatocellular marker ( Figure 1Z ). Case 4 was a beautiful example of hepatocellular carcinoma in the mouse and featured a liver section diffusely effaced and replaced by an infiltrative neoplasm composed of pleomorphic hepatocytes arranged in thick cords and trabeculae and containing large areas of hemorrhage and necrosis ( Figures 1O – 1Q ). Neoplastic cells exhibited a high mitotic index, hyperchromatic nuclei and marked anisocytosis/anisokaryosis ( Figure 1Q ). There was also a distinct focus of haphazardly arranged vascular channels coursing through the hepatocellular neoplasm and disrupting the continuity of the trabecular pattern within the carcinoma ( Figure 1R ). The diagnosis of hepatocellular carcinoma was further confirmed with negative CD31 ( Figure 1U ) and positive CK18 ( Figure 1W ) IHC staining. However, the focus of haphazardly arranged vascular channels within the hepatocellular carcinoma stained positive for CD31 ( Figure 1V ) and negative for CK18 ( Figure 1X ), confirming Dr. Quist’s suspicions that this was a metastatic hemangiosarcoma that likely originated from the hemangiosarcoma presented as Case 3. After voting for Case 4 was completed, Dr. Quist revealed that the reason why she believed that this focus was likely to be a metastatic hemangiosarcoma within the hepatocellular carcinoma was because she knew that Cases 3 and 4 were from the same animal, which she had deliberately not shared prior to voting. For the discussion, Dr. Quist reviewed key diagnostic features of liver tumors in the mouse as summarized in the International Harmonization of Nomenclature and Diagnostic Criteria (INHAND) guidance document ( Table 2 ). 1 Dr. Quist further summarized diagnostic take home points by emphasizing the importance of recognizing growth pattern and behavior when differentiating hepatic neoplasms in the mouse. As demonstrated in the case examples, growth patterns of hemangioma/hemangiosarcoma are characterized by irregular vascular channels, hepatocellular carcinoma typically exhibits a growth pattern of thickened trabeculae, and hepatoblastoma is a neoplasm comprised of small, intensely basophilic and elongated cells radially arranged around a central vascular channel, forming rosettes, trabeculae or pseudo-glandular structures. As for behavior, the location and appearance of metastasis can also be helpful in differentiating liver tumors in the mouse. Hepatic hemangiosarcomas readily metastasize to other parts of the liver or spleen and are typically well-differentiated. Hepatocellular carcinomas and hepatoblastomas also frequently metastasize but most often to the lung and are poorly differentiated. Dr. Quist also revisited the usefulness of IHC as a diagnostic tool for liver tumors in the mouse. For vascular neoplasms like hemangioma and hemangiosarcoma, CD31 and Factor 8 are helpful endothelial cell markers. For the liver, CK18 and CK19 are often used to confirm hepatocellular origin. However, hepatoblastoma is very elusive when it comes to IHC and does not stain positive with any known hepatocellular markers. A guidance document published by the American College of Pathologists 2 suggests using integrase interactor I (INI1), Glypican-3, and beta-catenin for hepatoblastoma but, as one audience member, and IHC expert, confirmed during the audience discussion, there are currently no reliable IHC markers for hepatoblastoma.

Dr. Martha Hensel (University of Texas MD Anderson Cancer Center [MDACC], Bastrop, Texas) presented two interesting cases of spontaneous tumors from nonhuman primate colonies at the Michael E Keeling Center for Comparative Medicine and Research. The animals were experimentally naïve and part of an inhouse breeding program. Dr. Carolyn Hodo (MDACC) was the primary pathologist on both cases and contributed to the presentation. The author would like to acknowledge Drs. Stanton Gray and Stephanie Buchl for providing the clinical perspective for the cases. The cases presented included colon from a rhesus macaque ( Macaca mulatta ) and cervix from an owl monkey ( Aotus nancymaae ). For the two cases, a series of low and high magnification photomicrographs were presented to the audience followed by voting choices. Case 1 featured colon from a 22-year-old, multiparous, female rhesus macaque ( M. mulatta ). This animal was heterozygous for a missense mutation in MSH6 , one of the mismatch repair genes associated with hereditary nonpolyposis colorectal cancer or Lynch syndrome. 27 Clinical findings included weight loss, positive fecal occult blood, and a palpable mass in the caudal abdomen. At necropsy, the abdominal cavity contained approximately 20 ml of clotted blood. Attached to the mesentery in the caudal left abdomen was a slightly firm, 4x3x2 cm, mottled, dark purple and tan mass. The transverse colon had an approximately 4cm diameter, infiltrative mural mass that expanded the wall up to 5 mm thick, and the mucosa was irregularly roughened with blood-stained contents. The surrounding mesentery was thickened and fibrous, and the mesentery of the distal jejunum was incorporated into the fibrous tissue (scirrhous response). A series of low and high magnification hematoxylin and eosin (HE)-stained sections of the mass in the colon were presented to the audience for their diagnosis ( Fig 5 A - D ). The low magnification image ( Fig 5A ) provided orientation of the lesion and highlighted the invasive nature of the proliferative glandular tissue that extended from the serosa to the submucosa. Higher magnification images of the region of interest in the muscular wall of the colon demonstrated that the lesion had degrees of morphological variation ( Fig. 5B ). The predominant lesion was composed of characteristic invasive endometrial glands lined by cuboidal to tall columnar endometrial cells supported by stroma; glands were surrounded by and contained blood ( Fig 5C ). A second lesion was composed of cuboidal cells arranged in nests and tubules on a desmoplastic stroma ( Fig 5 D ). Nuclei were irregularly round with finely stippled chromatin and 1 small nucleolus. Anisocytosis was moderate; no mitotic figures were noted in the section available for review. The audience was polled for their preferred diagnosis. Voting choices with voting results were as follows: endometrial carcinoma (10%), colon adenocarcinoma (29%), endometriosis (49%), retroperitoneal fibromatosis (11%), or other (1%). The audience concurred with the diagnosis preferred by the speaker: endometriosis. However, this case was selected because it was not considered a straightforward diagnosis. At necropsy the featured section of colon was originally diagnosed as a colon carcinoma due to the gross appearance and supporting evidence of a known genetic predisposition to developing colorectal cancer plus a positive fecal occult blood test. Endometriosis was diagnosed based on histologic examination, but the second morphologic population with the scirrhous stroma supporting nests and tubules of cuboidal epithelium suggested an additional diagnosis of either colon carcinoma or endometrial carcinoma. Endometriosis is a common reproductive tract lesion in aged female rhesus macaques with a prevalence approaching 30% in some populations. 28 , 29 The incidence may increase in colonies with a high rate of Caesarean section deliveries as one of the proposed pathogeneses is explantation of endometrial tissue during surgery. 30 Endometriosis is also diagnosed in up to 10% of reproductive aged women, and rhesus macaques can serve as an animal model for the condition. Characteristic histologic features include glandular epithelium and stroma that mimic endometrium. The glands may be filled with blood and/or degenerate and viable macrophages. Hemorrhage may also extend into the stroma and be accompanied by macrophages with erythrocytes or hemosiderin pigment. The lesion can wax and wane with the estrous cycle and induce hemoabdomen if the cysts rupture. Endometriosis is most common in the ovary, but ectopic endometrial tissue can be found throughout the abdomen including the colon. Endometriosis often appears as variably sized, brown, nodular lesions along the omentum, mesentery, or serosal surface of the organs and are known colloquially as “chocolate cysts.” Colon carcinomas may arise from the mucosa of the ascending, transverse, or descending colon as solitary or multifocal nodular masses with an accompanying scirrhous response that often causes constriction of the lumen. 31 A comparison of the typical appearance of colon carcinoma and endometriosis was presented to highlight the gross features of each condition. As part of the signalment, the presenter offered the information that the animal was heterozygous for MSH6 , one of the four genes associated with hereditary nonpolyposis colorectal cancer or Lynch syndrome in humans. Germline mutations in mismatch repair genes MLH1, MSH2, MSH6 , or PMS2 result in microsatellite instability that predispose humans to developing colorectal and/or endometrial cancer. 32 Mutations in the MLH1 and MSH6 have been identified in rhesus macaques that develop colorectal carcinoma at the Keeling Center; endometrial carcinoma has not yet been identified in our colony in association with one of the aforementioned genetic mutations. 27 The impact of mutations on development of colorectal disease was discussed. Animals in the colony with identified mutations appear to develop colorectal cancer at a younger age than the spontaneous cases although this has not yet been thoroughly investigated. Unlike mutations in humans, the influence of genetic predisposition on incidence in rhesus macaques is a subject for future research. Case 2 was tissue from a 4-year-old female owl monkey ( Aotus nancymaae ) that had received a Caesarean section for a term pregnancy. At the time of surgery, the animal had an enlarged urinary bladder. After surgery the animal was not urinating, and the veterinarian attempted manual expression but was unsuccessful, so euthanasia was elected. A mass was not noted in the urinary tract at necropsy. However, the pelvic canal contained several tortuous congested blood vessels, and the cervix appeared thickened. The mucosal surface of the cervix was thickened by a multilobulated mass. On cut section, the mass was mottled tan to red with depressed areas interpreted as foci of necrosis. A series of H&E-stained sections of the cervix were presented to the audience for their diagnosis ( Fig 5 E - H ). The low magnification image provided the anatomical orientation for the lesion as the location was not provided for the audience to increase the difficulty of the exercise ( Fig 5E ). However, Dr Hensel did point out the normal tissue on the left side of the image to help orient the viewers and offered the useful adage that if you are unsure of what you are looking at then you are probably in the reproductive tract. The second image highlighted the invasive nature of irregular tubules on a moderate fibrovascular stroma with scattered areas of hemorrhage ( Fig 5F ). Most neoplastic cells were cuboidal to columnar with variably distinct cell borders, a small amount of eosinophilic cytoplasm, and central oval nuclei with coarsely stippled chromatin and one prominent basophilic nucleolus ( Fig 5G ). Multifocally throughout the neoplasm was a second population of large multinucleated syncytial cells with hyperchromatic nuclei, which resembled syncytiotrophoblasts ( Fig 5G - H ). Neoplastic tubules often contained karyorrhectic and proteinaceous debris, and the stroma was multifocally hyalinized. After the images were reviewed, the audience was polled for their diagnosis. Voting choices with voting results were as follows: urothelial carcinoma (15%), gestational choriocarcinoma (20%), cervical adenocarcinoma (20%), epithelioid trophoblastic tumor (42%), or other (3%). Perhaps reflecting the difficulty of this case or the uncertainty of the anatomical location, fewer respondents weighed in with their selections (60 responses for case 2 compared to 68 for case 1). Nonetheless, 42% of respondents selected epithelioid trophoblastic tumor, which was the diagnosis preferred by Dr Hensel. The differentials considered included: epithelioid trophoblastic tumor, gestational choriocarcinoma, or cervical adenocarcinoma. Diagnostic features of each differential are presented in Table 6 . An epithelioid trophoblastic tumor was considered the most likely diagnosis due to the bland appearance of the cuboidal population of neoplastic cells and the presence of multinucleated cells with hyperchromatic nuclei. A gestational choriocarcinoma was considered less likely because these tumors reportedly have features of malignancy such as high mitotic rate, infiltrative and destructive growth pattern, large areas of necrosis, and intravascular invasion. None of these features were noted in the tumor presented. Cervical adenocarcinoma was ruled out due to the presence of the multinucleated cells, which are not a feature of this tumor. To confound diagnostic confidence, there is overlap between the trophoblastic tumor types, and epithelioid trophoblastic tumors may be part of a mixed trophoblastic tumor. 33 However, our diagnostic confidence was increased by the concurrence of our fellow pathologists at the meeting. This case was an interesting example of an epithelioid trophoblastic tumor, which is a rare diagnosis in nonhuman primates; our literature search turned up a single case report in the ovary of a cynomolgus macaque ( M. fascicularis ). 34 Overall, this owl monkey had a rare tumor type in an uncommon location; neoplasia of the cervix is not commonly reported in non-human primates and not at all in owl monkeys. 28 , 35 Audience members questioned why the animal could not urinate following delivery. Compression of the urethra by the neoplastic tissues possibly exacerbated by swelling associated with labor prior to C-section is considered the most likely explanation for the inability to urinate. After the presentation was concluded, an audience member offered a hypothesis for why the clinical signs only appeared after parturition: perhaps the weight of the gravid uterus exerted enough tension to keep the urethra open despite the space occupying nature of the cervical mass. Once the tension was alleviated (i.e., the infant was delivered), the counterbalance was removed, and the mass effect was observed.