Leukemic Cells Infiltrate the Ovaries Without Damaging Ovarian Reserve in an Acute Myeloid Leukemia Mouse Model

Abstract

Leukemia is one of the most common cancers in prepubertal girls and adolescents, with advances improving survival rates. However, treatments like chemotherapy and radiation are highly gonadotoxic, often causing ovarian insufficiency, early menopause, infertility, and endocrine disorders. Fertility preservation for young female patients with cancer, especially prepubertal girls without mature germ cells, relies heavily on ovarian tissue cryopreservation. Yet, a major concern is the potential presence of leukemic cells within preserved tissue, posing a risk of reintroducing malignancy upon grafting. Additionally, the direct effects of leukemia on ovarian function remain unclear. In this study, we used an acute myeloid leukemia (AML) mouse model to explore the impact of leukemia on ovarian function. Leukemic cells infiltrated the ovaries, particularly the stromal regions and granulosa layers of antral follicles, while also being present in the spleen and liver. Despite this infiltration, ovarian structure, follicular counts, and primordial follicle reserves were largely preserved, with the notable absence of corpus luteum indicating impaired ovulation. Furthermore, leukemic infiltration induced inflammatory cytokines TNF-α and COX-2, potentially influencing ovarian health. These findings suggest opportunities for fertility preservation by selectively removing leukemic cells, though risks of malignancy remain. This model offers a platform for advancing fertility-preservation strategies during gonadotoxic cancer therapies.

Leukemia is a malignancy of the immature hematopoietic cells of the bone marrow. It is one of the most frequently diagnosed cancers among individuals under the age of 40. The overall 5-year survival rate of leukemia increased to 88% for children younger than 15 years in 2024 (1). Acute myeloid leukemia (AML) is the most common acute leukemia in adults, accounting for about 1% of all cancer diagnoses and approximately 15% to 20% of cases in patients aged 15 years or younger (2). Leukemia typically arises in the bone marrow, and then progresses to infiltrate various organs, including liver, spleen, and lymph nodes, leading to common manifestations such as splenomegaly, hepatomegaly, and lymph node enlargement, respectively (3-7).

Women are born with a finite number of primordial follicles that make up their ovarian reserve. The nonrenewable reserve serves as the source of maturing follicles to sustain female reproductive cycles and fertility. The number of primordial follicles determines a woman's reproductive lifespan, and the depletion of these follicles leads to menopause (8-10). Cancer therapy affects the ovary through various mechanisms, making oocytes, the female germ cells, and granulosa cells susceptible to gonadotoxic anticancer agents. This can lead to the loss of primordial follicles and result in primary ovarian insufficiency (POI). POI can cause a range of reproductive and systemic issues, including infertility, sexual dysfunction, decreased bone density, increased cardiovascular risk, depressive and anxiety symptoms, and early mortality (11-14). Additionally, cancer cells can directly impact fertility, especially when tumors involve the reproductive organs (15). Various mechanisms have been proposed to explain the detrimental effects of cancer on fertility, including systemic response such as cytokine release from cancer cells, the body's immune response, and the stress associated with a cancer diagnosis, all of which can interfere with hormonal balance and impair fertility (16, 17). However, the exact mechanisms remain unclear. Despite recent advancements in novel therapeutic strategies and targeted drugs that have significantly improved the overall survival rate of leukemia, both leukemia itself and related gonadotoxic treatments have been suggested to have adverse effects on fertility (18, 19).

Given the significant impact of POI on quality of life, ovarian tissue cryopreservation has emerged as a viable option for preserving the fertility of young female patients with cancer (20, 21). Recently, the American Society for Reproductive Medicine guidelines state that “ovarian tissue banking is an acceptable fertility preservation technique and is no longer considered experimental. It is the only method to preserve fertility for prepubertal girls, as ovarian stimulation and IVF are not options” (22). However, it remains uncertain whether ovarian follicles affected by leukemia can be preserved and function effectively in the future, due to the potential risk of reintroducing leukemic cells (23, 24). Therefore, the current study aims to develop an AML model in young adult female mice and investigate the effects of leukemia on the ovary. To the best of our knowledge, this is the first report to demonstrate the direct effects of leukemic cells on the ovary and primordial follicles using a leukemic mouse model.

Materials and Methods

Animals

All animals used and procedures performed in this study were approved by the Institutional Animal Care and Use Committee at the University of Nebraska Medical Center (UNMC). The animals had ad libitum access to food and water and were housed in the Comparative Medicine facilities at UNMC. The Cbfb^+/56M1; Gt(ROSA)26Sor^{tm4(ACTB-tdTomato, −EGFP)Luo/SjJ}; Mx1-Cre⁺ (Cbfb^+/MYH11, GFP⁺, Mx1-Cre⁺) AML mouse model was bred in Dr. Hyde's laboratory as previously described (25-27). In brief, mice were treated with a 250 µg/dose of polyinosine–polycytidylic acid via intraperitoneal injection to induce Cbfb::MYH11 expression (Fig. S1A (28)). The knock-in mice developed leukemia at 2 to 3 months after induction and were sacrificed after leukemia development was confirmed. Peripheral blood cells were isolated from Cbfb^+/MYH11; GFP⁺; Mx1-Cre+ female mice and incubated in ACK lysis buffer (A1049201, ThermoFisher Scientific, Waltham, MA). Bone marrow cells were harvested from femurs and stained with fluorescein isothiocyanate-, phycoerythrin-, and allophycocyanin-conjugated antibodies against Csf2rb (CD131), c-Kit, and GFP (Fig. S2 (28)). Spleen and liver tissues from transgenic female mice were also collected to confirm leukemia development and stained for GFP⁺ cells (Figs. S1B-S1D (28)). After confirming the development of leukemia using fluorescence-activated cell sorting (FACS), GFP-positive leukemic cells were sorted by FACS. They were collected and stored at −80 °C.

FACS-sorted, GFP⁺ leukemic cells (5 × 10⁵) were transplanted into 1-month-old congenic (C57BL/6/129SvEv F1) mice via retro-orbital injection. The femurs of recipient mice were irradiated using the image-guided irradiation (IG-Rad) approach with the Small Animal Radiation Research Platform (Xstrahl, Suwanee, GA) to allow engraftment of leukemia cells in the bone marrow while protecting ovarian follicles from sublethal irradiation (630 cGy). This approach was chosen to avoid the total body irradiation that destroys ovarian follicles, particularly primordial follicles (29). IG-Rad is a superior method for preserving the majority of ovarian follicles in leukemic mouse models, although it still causes damage to primordial follicles compared with the ovaries of untreated mice (Fig. 1B). IG-Rad (5 Gy of X-rays) was performed when mice were 4 weeks old in Dr. Oberley-Deegan's laboratory. Briefly, the isocenter was set for the femur bones of both legs and 5 Gy of radiation was administered using a 15 × 15 mm collimator to minimize side effects on the gonads. Flow cytometry analyses were conducted on peripheral blood that was collected to monitor leukemia engraftment and progression. When the percentage of GFP⁺ leukemic cells in the peripheral blood averaged 20%, which was typically around 2 weeks after injecting 5 × 10⁵ GFP⁺, the mice were euthanized. The spleen and bone marrow were harvested to assess leukemic status, and the ovaries and uteri were collected to evaluate reproductive health.

Figure 1.

The leukemia mouse model retains primordial follicles in the ovary. (A) Schematic representation of the leukemia mouse model establishment. GFP-positive leukemic cells are isolated from Cbfb^+/MYH11; GFP⁺; Mx1-Cre⁺ mice and transplanted into 4-week-old female mice that are exposed to image-guided radiation. Approximately 2 weeks later, FACS analysis confirms leukemia development, and the ovaries are then isolated. (B) Total numbers of each class of ovarian follicles in control (Con, n = 3 mice), total body radiation (Rad, n = 3), and image-guided radiation (IG-Rad, n = 3) groups. Both ovaries from each mouse were included in the follicle count. ***P < .001; ****P < .0001; n.s., not significant. (C) Gross histological images of ovaries showing different follicle types: PF, primordial follicle; PM, primary follicle; SF, secondary follicle; CL, corpus luteum. (D) Total numbers of primordial follicles per ovary in the IG-Rad (n = 3 mice) and the leukemic mouse model (n = 3 mice) groups. n.s., not significant. (E) Primordial follicles are indicated by arrowheads in both groups. (F) Growing follicles, including primary (PM), secondary (SF), and antral follicles (AF) are indicated by arrowheads in both groups. (G) Total numbers of each follicle class per ovary in the IG-Rad (n = 3 mice) and the leukemic mouse model (n = 3 mice) groups. n.s., not significant; ***P < .001.

Follicle Counting

Ovaries were collected from female mice and fixed in Modified Davidson's Fixative (64133-50, Electron Microscopy Sciences, Hatfield, PA) for no more than 24 hours at 4 °C. The tissues were then processed, embedded in paraffin, and sectioned at a thickness of 5 μm. Sections were stained with hematoxylin and eosin (H&E) for histological analysis. Counting and categorization of each class of ovarian follicles were performed as previously described (29-33). Primordial follicles, characterized by an oocyte surrounded by a single layer of squamous pregranulosa cells, were counted on every tenth slide throughout the ovary, provided they had discernible nuclei covered with pregranulosa cells. Primary follicles, featuring an oocyte surrounded by a single layer of cuboidal granulosa cells, were similarly identified. Secondary follicles, defined by an oocyte surrounded by multiple layers of cuboidal granulosa cells, and antral follicles, marked by the presence of an antrum of any size, were counted when they exhibited visible nucleoli and granulosa cells in H&E-stained sections. To calculate the total number of primordial and primary follicles per ovary, the average follicle count per section was multiplied by the total number of sections and then divided by 2, accounting for the approximate 10-μm nucleus span requiring at least 2 discernible sections. For secondary and antral follicles, the total count per ovary was obtained by multiplying the average follicle count per section by the total number of sections.

Histology, Immunofluorescence, and Immunohistochemistry Assays

H&E staining, immunofluorescence, and immunohistochemistry were performed as previously described (29, 33-36). For immunofluorescence assays, the ABC-HRP kit (PK-6100, Vector Laboratories, Inc., Newark, CA) and Alexa Fluor 488 Tyramide SuperBoost kit (B40932, ThermoFisher Scientific) were used, while the Metal Enhanced DAB Substrate kit (34065, ThermoFisher Scientific) was utilized for DAB staining. The catalog numbers and dilution of primary antibodies were as follows: α-SMA (1924, RRID:AB_2223021, 1:50), tumor necrosis factor (TNF)-α (6945, RRID:AB_10859375, 1:50), cleaved CASP3 (9661, RRID:AB_2341188, 1:100), TAp63α (13109, RRID:AB_2637091, 1:100) from Cell Signaling; GFP (15256, RRID:AB_10979281, 1:50) from Invitrogen; fluorescent cKit (CD117, RRID:AB_400044), and CSF2RB (CD131, RRID:AB_2738658) for flow cytometry from BD Biosciences (Franklin Lakes, NJ), CD31/PECAM-1 (AF3628, RRID:AB_2161028, 1:100) from R&D systems (Minneapolis, MN); DDX4 (ab270534, 1:100) from Abcam (Cambridge, UK), anti-Müllerian hormone (AMH) (sc-6886, RRID:AB_649207, 1:50) and cyclooxygenase (COX)-2 (376861, RRID:AB_2722522, 1:50) from Santa Cruz Biotechnology, Inc. (Dallas, TX), α-Laminin (L9393, RRID:AB_477163, 1:100) from Sigma-Aldrich (St. Louis, MO); Inhibin α (1:100) was kindly provided by Dr. Wylie Vale, Salk Institute (La Jolla, CA). Secondary antibodies used were as follows: goat antirabbit IgG (H + L) (BA-1000, RRID:AB_2313606, 1:400) and goat antimouse IgG (H + L) (BA-9200, RRID:AB_2336171, 1:400) from Vector Laboratories, Inc. The specificity of secondary antibody binding was confirmed by the absence of signal in negative controls incubated without primary antibodies. All images were captured using the EVOS M7000 Imaging System (AMF700, Invitrogen, Carlsbad, CA). To quantify the signals from immunofluorescence assay, we used ImageJ. TNF-α and COX-2 intensities were measured from n = 3 samples in the IG-Rad group and n = 4 samples in the leukemia group, with 3 to 10 images analyzed per sample. The pixel intensities were then averaged.

Picrosirius Staining

Fibrosis in the ovaries was assessed using the Picrosirius Red stain kit (MER PSR1, Mercedes Scientific, Lakewood Ranch, FL) following the manufacturer's protocol. Ovarian fibrosis was quantified as the collagen proportionate area, defined as the ratio of the area occupied by Picrosirius Red-stained collagen fibers to the total area of the section. The analysis was performed at a high magnification of 40×. To assess the reproducibility of the quantification, the analysis in ImageJ was repeated 3 times.

TdT-Mediated dUTP Nick-End Labeling

Apoptosis was evaluated by detecting DNA fragmentation using the DeadEnd fluorometric TdT-mediated dUTP Nick-End Labeling (TUNEL) system (G3250, Promega) in accordance with the manufacturer's instructions.

Statistical Analysis

Graphs were generated using Prism 10.2.3 software (GraphPad Software version 10), and data were presented as means ± SEM. One-way analysis of variance with Tukey's post hoc test was used to determine statistical differences among groups. An unpaired 2-tailed Student t-test was used for pairwise comparisons between groups. P values of less than .05 were considered statistically significant.

Result

The Leukemic Mice Exhibited an Ovarian Reserve Comparable With That of Noncancerous Controls

To generate leukemia cells, knock-in mice expressing the Cbfb::MYH11 fusion oncogene from the endogenous Cbfb locus (Cbfb^+/MYH11) were crossed with transgenic mice expressing Cre recombinase under the inducible hematopoietic Mx-1 promoter, and GFP from the Rosa26 locus (Fig. S1A (28)). Six- to 10-week-old Cbfb^+/MYH11; GFP⁺; Mx1-Cre⁺ mice were induced to express Cbfb::MYH11 and monitored for leukemia development by assessing the percentage of malignant cells in peripheral blood using GFP, and leukemia markers cKIT and CD131 (Fig. S2 (28)). Upon sacrifice, Cbfb^+/MYH11; GFP⁺; Mx1-Cre+ mice (Fig. S1A (28)) displayed splenomegaly and hepatomegaly (Fig. S1B, with green arrows indicating spleens and yellow arrows indicating livers) (28). Histological images of the spleen and liver revealed increased cell density (Fig. S1C) (28), and immunofluorescent microscopy confirmed leukemia development, showing infiltrated GFP-positive leukemic cells with large nuclei (Fig. S1D and S1E) (28).

In order to model leukemia development in pediatric patients, GFP⁺ leukemic cells were transplanted into 1-month-old wild-type congenic (C57BL/6/129SvEv F1) mice (Fig. 1A). To facilitate leukemia cell engraftment, bone marrow myeloablation is typically required, achieved through total body irradiation or pharmacological methods (Fig. 1A). However, these methods have been shown to impair ovarian function. IG-Rad potentially provides sufficient myeloablation for leukemia cell engraftment while sparing ovarian tissue. To test this possibility, we assessed the impact of irradiation dose on ovarian follicles and compared the number of surviving ovarian follicles across control (Con), whole-body radiation (Rad), and IG-Rad groups. Follicle-counting data revealed the presence of primordial follicles in the IG-Rad group, with significantly higher numbers of follicles at all stages, including primordial follicles, than the total body irradiation group (Fig. 1B). While IG-Rad reduced primordial follicle numbers compared with controls, it still offered significant advantage over whole-body radiation by minimizing inevitable ovarian damage.

Leukemic mice, approximately 6-8 weeks old, were euthanized upon reaching 20% leukemia in peripheral blood (Fig. S2 (28)). Ovarian morphology was then assessed in leukemia mice transplanted with GFP⁺ leukemic cells and in IG-Rad–only mice exposed to IG-Rad without leukemic cell transplantation. Compared with IG-Rad–only mice, leukemic mice showed similar morphology across all follicle stages (Fig. 1C) and comparable primordial follicle counts (Fig. 1D). The total number of primordial follicles in leukemic mice was comparable to that of the control group (Fig. 1D). The histological image shows the presence of primordial follicles in both the leukemia and IG-Rad groups (Fig. 1E). Additionally, primary, secondary, and antral follicles were observed in the ovaries of leukemic mice, and their numbers were comparable to those in the control group (Fig. 1F and 1G). However, none of the leukemic mice exhibited a corpus luteum, an indicator of ovulation (Fig. 1C and 1G).

Leukemic Cells Infiltrate the Ovary

Leukemic cells are known for their ability to infiltrate various organs via systemic circulation, yet there is limited research on their effects on the ovary. To investigate whether leukemic cells penetrate the ovary and the potential ovarian impact, we examined the ovaries of leukemic mice using a GFP antibody. The results revealed the presence of GFP⁺ leukemic cells within the ovaries of leukemic (IG-Rad + leukemic cells) mice (Fig. 2A), predominantly localized in the stroma area surround growing follicles, including primary, secondary, and antral follicles (Fig. 2A, red arrows). Additionally, GFP⁺ leukemic cells were randomly distributed throughout the ovary of leukemia mice (Fig. 2B) and were also observed in the uterus (Fig. S3 (28)).

Figure 2.

Leukemic cells penetrate the mouse ovary without damaging ovarian structure. (A) Immunofluorescence images of GFP expression in whole ovarian sections from 8-week-old control mice (IG-Rad, top) and leukemia mice (bottom). Nuclei are stained with DAPI (blue). GFP⁺ leukemic cells are indicated by green fluorescent signals, marked with arrows. Nonspecific background staining is denoted by asterisks. (B) High-magnification images from immunofluorescence assays. (C) Immunofluorescence images of markers: laminin for basement membrane, α-SMA for theca cells, and CD31 for endothelial cells. (D) Expression of TAp63α and DDX4 for oocytes, and AMH for granulosa cells.

To evaluate whether ovarian structure was preserved throughout the ovary of leukemia mice, we performed immunofluorescent assays using 3 established markers: laminin, a key component of the basement membrane; α-SMA, a component of the blood vessel walls, muscle fibers, and stromal cells surrounding the follicles; and CD31, an endothelial marker. The localization patterns of these structural markers were similar between the control (IG-Rad only) and leukemic (IG-Rad + leukemic cells) groups (Fig. 2C), implying that ovarian structure was maintained despite the infiltration of leukemic cells.

We also assessed the expression of oocyte markers TAp63α and DDX4. TAp63α is mainly expressed in the oocytes of primordial follicles, while DDX4 is present in the cytoplasm of oocytes from primordial germ cells to preovulatory follicles. The expression of these markers indicates that oocytes in leukemic mice maintain normal expression despite leukemic cell infiltration (Fig. 2D). Additionally, the expression of AMH, a member of the transforming growth factor-β superfamily produced by granulosa cells of growing follicles, including preantral and small antral follicles, confirmed that granulosa cells in leukemic ovaries maintained their molecular signatures despite the presence of leukemic cells (Fig. 2D).

Leukemic Cells Penetrate Antral Follicles

To determine whether leukemic cells penetrate into the ovary and enter the theca layer, granulosa cell layers, and follicular fluid of antral follicles, we examined the localization of GFP⁺ cells in ovarian sections from IG-Rad only and leukemia mice. GFP⁺ leukemic cells (orange arrowheads) were predominantly observed in areas of the stromal cells surrounding growing follicles, including primary, secondary, and antral follicles (Fig. 3A). Notably, GFP⁺ leukemic cells were also found within several antral follicles (red-dotted lines), with some cells localized near the theca layers and others dispersed throughout the granulosa cell layers (Fig. 3A). Interestingly, some antral follicles contained GFP⁺ leukemic cells in both the very inner and outer layers of mural granulosa cells, showing the presence of GFP⁺ leukemic cells among mural granulosa cells (Fig. 3B). However, primordial, primary, and secondary follicles did not contain GFP⁺ leukemic cells within the layers of granulosa cells (Fig. 3A). In addition, the stromal area surrounding primordial follicles (Fig. 3A, white arrowheads) did not show signs of leukemic cell contamination. The presence of GFP⁺ leukemic cells within the theca layer or stromal cell regions was further confirmed by costaining GFP with a marker for smooth muscle cell and for the theca layer of the ovary. This analysis revealed that GFP + leukemic cells were localized within the theca layers and spread into the granulosa cells (Fig. 3C, orange arrowheads). Notably, GFP⁺ leukemic cells were confined to the theca layers and did not intermingle with α-SMA-positive cells.

Figure 3.

Leukemic cells penetrate antral follicles. (A) Immunofluorescence images of GFP expression to detect GFP-positive leukemic cells (small arrowheads) in ovarian sections from control (IG-Rad) and leukemia mice. PF, primordial follicles (white dotted lines with big arrowheads); PM, primary follicles; SF, secondary follicles; AF, antral follicles. DAPI (blue) highlights the nuclei. (B and C) Double immunofluorescence assays of inhibin α and GFP (B) and α-SMA and GFP (C). GFP-positive leukemic cells (red) are indicated by small arrowheads. High magnification images are shown in a and b in (B). Leukemic cells are indicated by orange arrowheads (a and b), with high-magnification images shown below.

Primordial Follicles Are Not Affected by Leukemic Cell-Driven Inflammatory Factors

Tumor cells, including leukemia cells, release various growth factors and cytokines, such as TNF-α and COX-2, to evade immune surveillance and create a proinflammatory environment (37). To determine whether the infiltrated leukemia cells can use a proinflammatory environment in the ovary, we performed immunofluorescence assays to examine the expression of TNF-α and COX-2. Compared with ovaries from noncancerous controls (IG-Rad only), ovaries from leukemic mice exhibited significantly higher expression levels of TNF-α and COX-2 in the area of stromal cells surrounding growing follicles, including primary, secondary, and antral follicles but not in primordial follicles (Fig. 4A and 4B). To examine whether the expression of TNF-α and COX-2 affects the survival of ovarian follicles, particularly granulosa cells of antral follicles, TUNEL staining was performed. The results showed that some granulosa cells in antral follicles were TUNEL positive in the IG-Rad–only group, a common feature observed in normal mouse ovary. In contrast, multiple preantral and antral follicles were TUNEL positive in the leukemia group (Fig. 4C). This suggests that leukemic cells may create an inflammatory environment that contributes to granulosa cell death (Fig. 4C). Given that ovarian stroma inflammation can promote follicle atresia by inducing apoptosis of granulosa cells, we further investigated the expression of cleaved CASP3, a marker of apoptosis. Immunofluorescence staining revealed expected positive cleaved CASP3 signals in granulosa cells surrounding the antrum of antral follicles in the IG-Rad group. In contrast, numerous positive cleaved CASP3 signals were observed in various regions of multiple antral follicles, as well as in stromal cells, in the leukemia groups (Fig. 4D). However, no cleaved CASP3 signals were detected in primordial follicles in either group (Fig. 4E and 4F). This suggests that while granulosa cells in antral follicles are susceptible to apoptosis in both groups, primordial follicles appear to remain unaffected.

Figure 4.

Leukemic cells induce inflammatory events and apoptosis in the ovary. (A) Immunofluorescence analysis of TNF-α and COX-2 expression in ovarian sections. DAPI (blue) highlights the nuclei. (B) Quantification of TNF-α and COX2 expression in the IG-Rad group (n = 3) and the leukemia group (n = 4). (C) TUNEL assay of ovarian sections from IG-Rad and leukemia mice. (D) Expression of cleaved caspase 3 (cCas3, green) in ovarian sections from IG-Rad and leukemia mice. (E) Expression of cleaved caspase 3 (cCas3, green) in antral follicles of ovarian sections from IG-Rad and leukemia mice. Theca cell layers are outlined with red dotted lines. (F) Expression of cleaved caspase 3 (cCas3, green) in primordial follicles within the cortical area. GC, granulosa cells; PF, primordial follicles (white dotted lines); AF, antral follicles. DAPI (blue) highlights the nuclei. *P < .05.

Discussion

Multiple mouse models with hematologic malignancies have been developed to understand human hematopoiesis (38). However, these models are suboptimal for examining the effects of leukemic cancer cells on the ovary. Although the knock-in female mouse model effectively develops leukemia and mirrors the human leukemic process, there is significant latency before leukemia onset. This delay poses challenges for studying ovarian reserve and follicle development due to the onset of reproductive aging.

An alternative approach for developing a leukemia mouse model involves transplanting leukemic cells into congenic female mice. However, this method requires exposing the host females to high levels of radiation to eliminate hematopoietic cells in the bone marrow, which can also destroy ovarian follicles due to their sensitivity to radiation. To protect ovarian follicles from radiation damage, we used image-guided radiation in our study. Leukemia cells isolated from Cbfb^+/MYH11; GFP⁺; Mx1-Cre⁺ mice were transplanted into partially irradiated 1-month-old C57BL/6/129SvEv F1 mice to induce leukemogenesis.

GFP⁺ leukemic cells were detected in the liver and spleen of the transgenic leukemic mouse model, with these organs significantly enlarged compared with those of the control group. This indicates leukemic cell infiltration into these organs, resulting in their abnormal enlargement. Additionally, this transgenic leukemic mouse model demonstrated evidence of leukemic cell infiltration into the ovaries, with leukemic cells identified in preantral and antral follicles, as well as in stromal cells. To our knowledge, this is the first report investigating the effects of leukemic cell infiltration into the ovary using a mouse model. Notably, although leukemic cells were observed throughout the ovary, the structural markers for oocytes, granulosa cells, theca cells, basement membrane, and vessels were expressed normally, comparable to the control ovary.

However, the mechanism by which leukemic cells penetrate ovarian follicles remains unclear. Previous studies have described the blood–follicle barrier (BFB) as a structural interface between blood vessels and ovarian follicles. This barrier comprises the vascular endothelium, subendothelial basement membrane, thecal interstitium, follicular basement membrane, and membrana granulosa (39-41). This barrier functions as a molecular sieve, permitting the permeation of low molecular weight proteins (<500 kDa). While proteins of medium or high molecular weight can cross the blood vessels into the interstitium, the follicular basement membranes restrict the permeability of these larger molecules. Beyond the BFB, ovarian follicles contain tight junction proteins (occludin, cingulin, claudins, and tight junction protein 1) (42), as well as adherens junctions and desmosomes (43), which further block the passage of even smaller molecules. A recent study by Kanatsu-Shinohara demonstrated the potential of gene delivery using adeno-associated virus to cross the BFB and restore oogenesis in congenitally infertile mice (44, 45). This suggests that leukemic cells may also be able to cross the BFB and penetrate ovarian follicles through mechanisms that are not yet fully understood.

One possible explanation for this observation is the structural remodeling of the ovary during follicular growth, particularly during antrum formation. In our mouse model, leukemic cells were not only present in stromal cells but also dispersed randomly among granulosa cells. Additionally, some leukemic cells were found in the inner cell layers of antral follicles. The theca layer undergoes rapid changes in its vascular supply in response to the ovulatory luteinizing hormone (LH) signal. Vascular endothelial growth factor and its downstream signaling pathways are crucial for follicle angiogenesis during antral follicle growth. Vascular endothelial growth factor enhances vascular permeability, which facilitates the efficient delivery of factors such as LH, follicle-stimulating hormone, and immune cells to the follicle. This increased permeability may potentially allow leukemic cells to penetrate the antral follicle.

The ovaries of the leukemic mice in this study exhibited elevated levels of TNF-α and COX-2 in stromal cells. Both TNF-α and COX-2 are known to play critical roles in leukemogenesis, contributing to cellular transformation and proliferation (46, 47). Notably, TNF-α can upregulate COX-2 expression through specific intracellular signaling pathways (48).

Leukemic cells are well-documented producers of proinflammatory cytokines such as TNF-α (49, 50), IL-1β (51, 52), CCL3 (53, 54), and CXCL2 (50), which collectively create a proinflammatory environment that supports the leukemic niche (55) and stimulates leukemic cell growth. This inflammatory milieu helps leukemic cells evade immune surveillance and enhances their survival, ultimately leading to poorer clinical outcomes (56). Additionally, leukemic cells from patients with AML have been shown to secrete IL-6 in vitro, which contributes to chemoresistance (57) and enhances growth factor-dependent proliferation of AML blasts (58). These findings suggest that inflammatory factors secreted from leukemic cells amplify inflammatory responses in the ovary. Literature indicates that these inflammatory factors can disrupt ovarian follicles by inhibiting ovulation and inducing apoptosis in granulosa cells, which leads to follicle depletion (59, 60). Thus, ovarian follicles may be adversely affected by hematologic malignancies if not addressed in a timely manner.

The most striking phenotype in this leukemic mouse model is the absence of a corpus luteum in the ovary, a feature not observed in the control group. One potential explanation is that the inflammatory environment created by leukemic cells induces the death of granulosa cells and all growing follicles, including primary, secondary, and antral follicles, ultimately preventing corpus luteum formation. This is supported by the presence of cleaved CASP3 and TUNEL-positive cells in the granulosa cells of growing and antral follicles.

Another possibility is that granulosa cells in the ovaries of leukemic mice are functionally impaired, leading to insufficient production of key hormones such as estrogen, progesterone, activin, and AMH. This hormonal deficiency could disrupt the communication between the ovary and the hypothalamus–pituitary axis, resulting in altered secretion of essential reproductive hormones like follicle-stimulating hormone and LH. A third hypothesis is that leukemic cells infiltrate the brain, particularly the hypothalamus or pituitary, disrupting the regulation and production of hormones critical for normal ovarian function. Therefore, further research is needed to investigate how the inflammatory milieu induced by leukemic cells affects folliculogenesis, ovarian function, and overall reproductive health.

This leukemic mouse model demonstrates the presence of intact primordial follicles in the ovary, without any apparent damage or infiltration by leukemic cells. However, the long-term health and viability of these follicles remain uncertain, particularly given their heightened sensitivity to stressors such as chemotherapy, environmental toxins, and inflammatory factors. Although primordial follicles appear to be unaffected in the short-term, their vulnerability under prolonged inflammatory or leukemic conditions warrants further investigation. Thus, preserving primordial follicles, rather than ovarian cortical tissue or growing follicles, including primary, secondary, and antral follicles, may be the most effective strategy for protecting ovarian function in patients with leukemia. This study demonstrates that this mouse model is valuable for investigating ovarian function in leukemic mice and provides important insights into the effects of leukemic cells on the ovary, particularly on primordial follicles.

Abbreviations

AMH

anti-Müllerian hormone

AML

acute myeloid leukemia

BFB

blood–follicle barrier

COX

cyclooxygenase

FACS

fluorescence-activated cell sorting

H&E

hematoxylin and eosin

IG-Rad

image-guided irradiation

LH

luteinizing hormone

POI

primary ovarian insufficiency

TNF

tumor necrosis factor

UNMC

University of Nebraska Medical Center

Funding

This work was supported by National Institutes of Health grant R01HD096042, R01HD115810, and startup funds from University of Nebraska Medical Center (UNMC) (S.-Y.K.), National Institutes of Health grant R01CA244900 (R.K.H), and National Institutes of Health grant S10OD023447 (R.E.O.-D.).

Author Contributions

A.A., Y.L., W.S., M.B., S.P., S.A.S., E.A.K., R.E.O.-D., S.X., R.K.H., and S.-Y.K. designed the study, performed the experiments, analyzed the results, drafted the manuscript, and revised the manuscript. A.A. and Y.L. performed most experiments. A.A., Y.L., W.S., R.E.O.-D., S.X., R.K.H., and S.-Y.K. revised the manuscript. A.A., Y.L., and S.-Y.K directly accessed and verified the underlying data reported in the manuscript. All authors have read and approved the final version of the manuscript.

Disclosures

The authors declare that they have no conflict of interests.

Data Availability

Original data generated and analyzed during this study are included in this published article.