Introduction

Caveolin-1 (CAV1) is located on chromosome 7q31.2 at the D7S522 locus, a fragile point known to be deleted in certain human cancers [1]. CAV1 is commonly described with a two-faced nature as both a tumor suppressor and an oncogene [2]. This paradox is seen across multiple malignancies as the expression of CAV1 has been shown to be both up- and down-regulated compared to normal tissue, with both phenotypes capable of harboring poor prognoses [2, 3].

With the “Janus-faced” nature of CAV1, researchers have turned to Cav1-deficient mice to sort out the heterogeneity. Cav1−/− mice, but not Cav1+/− mice, harbor a reduced lifespan in comparison to wild-type (WT) controls; however, the reduction in lifespan is not due to an increased frequency of tumorigenic events [4, 5]. Rather, pulmonary and cardiac complications drive the observed lifespan reduction [4]. It is clear that Cav1 plays a largely context-dependent role, with varied phenotypes seen throughout different organ systems as well as a progression of these phenotypes over time. Nonetheless, it is still unknown how Cav1-deficiency and heterozygosity would impact the development and composition of the primary and secondary lymphoid organs of mice. Cav1 is expressed in cells of the immune system [6, 7] and plays a vital role in not only immune synapse formation [8] but also in macrophage-, T-, and B-cell signaling [8,9,10,11,12,13,14]. While previous studies have investigated Cav1-deficiency in terms of B-cell development [10, 14, 15], T-cell development [10, 11], and recently the immunophenotype of the spleen (SP) and bone marrow (BM) in aged mice [14]; no comprehensive study which includes both primary and secondary lymphoid organs, a longitudinal component, and a Cav1-heterozygous experimental population exists.

Consequently, herein we investigated the impact of Cav1 on the lymphoid organ system. Specifically, we addressed how the primary and secondary lymphoid organs of mice can be immunophenotypically described at a young/non-diseased stage (12 weeks) and how the underlying pathologies of Cav1-deficiency impact these organs at near end-of-life (36 weeks).

Main text

Results

Caveolin-1-deficiency is associated with persistent splenomegaly

We observed no statistically significant differences in body weight (g) across genotypes and timepoints, with no sexual dimorphism (Fig. 1a, Additional file 1: Figure S1a). However, as early as 12 weeks, Cav1−/− mice presented with splenomegaly; a finding consistent across all measured timepoints in comparison to WT mice, with Cav1+/− mice mirroring WT mice and no observed sexual dimorphism (Fig. 1b, Additional file 1: Figure S1b).

Fig. 1
figure 1

Caveolin-1 deficiency is associated with splenomegaly with implications for Caveolin-1 as haploinsufficient for splenic B-cell maintenance. a Line graph of mean body weight (g) of Cav1+/+ (red), Cav1+/− (blue), and Cav1−/− (green) mice at indicated timepoint (n = 9–22 per group). b Line graph of mean SP weight (mg) of Cav1+/+ (red), Cav1+/− (blue), and Cav1−/− (green) mice at indicated timepoint (n = 9–15 per group). c (Top panel) Dot plots for CD19+CD22+ B cells in the SP for listed genotypes at 12, 24, and 36 weeks (n = 7–10 per group). (Bottom panel) Representative contour flow cytometry plots for CD19+CD22+ B cells at listed genotypes for listed timepoints. (Line graphs show mean ± SEM, dot plots show mean ± SD with each dot corresponding to a biological replicate, flow cytometry gate values show mean ± SEM, NS not significant, *p < 0.05 ANOVA and Tukey post hoc test)

Full size image

Caveolin-1 is dispensable for early lymphoid development but critical for B cell maintenance in the spleen

Investigating the impact of Cav1-deficiency and heterozygosity on B- and T-cell development and homeostasis, we observed no statistically significant differences in the pre/pro B-cell populations, immature B-cell populations, or overall CD19+ B-cell populations in the BM nor CD4/CD8 thymic populations in all experimental groups and time points (Additional file 2: Figure S2a, b). There were no observed changes in the resident CD3+ T-cell populations or skewing of the CD3+CD4+ versus CD3+CD8+ T-cell populations in the thymus (Thy), SP, lymph node (LN), or BM (Additional file 3: Figure S3a, b). Further, we observed no differences in the percentages of mature B cells in the LN or BM at 12 or 36 weeks (Additional file 3: Figure S3c).

Conversely, our aged Cav1−/− and Cav1+/− mice exhibited decreased CD19+CD22+ B-cell percentages in the SP as early as 24 weeks with corroborative results at 36 weeks (Fig. 1c). Our observation of a decrease in splenic B-cell percentages was seen without any concurrent changes to additionally analyzed white pulp populations, including: T cells (Additional file 3: Figure S3) or myeloid cells (Additional file 4: Figure S4a–c). With our flow cytometry controlled for cell input, events/sample (see “Methods”), and no change observed in overall lymphoid percentage (Additional file 4: Figure S4d), this excludes the splenomegaly of the Cav1-deficient mice as a confounding factor.

Aged Cav1-transgenic mice exhibit an altered splenic microenvironment

In response to the observed dysregulation of splenic B-cells in both Cav1−/− and Cav1+/−- mice, we conducted histopathologic analysis on splenic sections from all genotypes. Spleens from 12-week mice showed no differences under routine examination of H&E sections (Fig. 2a) with normal splenic architecture, including: similar frequency of observable secondary follicles (Fig. 2b) as well as mean secondary follicle area (Fig. 2c) in accordance with previous observations [10]. However, at 36 weeks, we noted Cav1−/− and Cav1+/− mice to present with fewer secondary follicles (Fig. 2a, b) with the observed follicles to be smaller in comparison to age-matched splenic sections from WT mice (Fig. 2c).

Fig. 2
figure 2

Aged Cav1-transgenic mice exhibit an altered splenic microenvironment. a Representative H&E stained sections of Cav1+/+, Cav1+/−, and Cav1−/− spleens at ×4 magnification. b Bar graph of the mean splenic secondary follicle area (μm2) for listed genotypes at 12 and 36 weeks (n = 5–10 follicles across n = 2–4 mice per group). c Percentage of observable secondary follicles for listed genotypes at 12 and 36 weeks (n = 2–4 per group). d Representative Ki67 and Movat’s Pentachrome stained histological sections from 36-week spleens for listed genotypes at 20× magnification (n = 3 per group). Increased fibrin staining is visualized in bright red. e Bar graph of mean plasma cells identified per field of view (×40) for listed genotypes from 36-week splenic sections (n = 3 per group). (Bar graphs show mean ± SD, with each dot corresponding to a biological replicate NS not significant, *p < 0.05 ANOVA and Tukey post hoc test)

Full size image

Additional investigation into the proliferative indexes of the 36-week sections revealed no statistically significant differences in Ki67+ cells per field of view (Fig. 2d). Utilizing pentachrome staining we noted an increase in fibrin deposits in the spleens of Cav1+/− and Cav1−/− mice compared to WT spleens (Fig. 2d). No other investigated organs were observed to harbor these deposits. Interestingly, further histopathologic analysis revealed an increase in inter-follicular plasma cells in the spleens of Cav1+/− and Cav1−/− mice compared to WT mice (Fig. 2e).

Follicle center B-cell lymphomas display altered CAV1 staining and expression patterns

Demonstrated above, we observed Cav1 to play a role in the long-term homeostasis of the splenic B-cell population and splenic follicle morphology. We therefore became interested in investigating the expression of CAV1 in post-secondary follicle B-cell malignancies. To this end we scored the Biomax LY6161 tissue microarray for CAV1 staining intensity and localization comparing healthy lymph node samples to both diffuse B-cell lymphoma (DBCL) and follicular lymphoma (FL) samples (Fig. 3a). The distribution of staining scores for disease entities differed in comparison to the normal LNs with fewer B-cell malignancies scored for heavy stromal staining and an increase in those scored for faint blood/lymphatic vessel staining (Fig. 3b). There was no significant difference in the distribution of staining scores when comparing malignancies. Using publicly available microarray data we examined the expression of CAV1 in tumor biopsy samples from FL and two established subtypes of diffuse large B-cell lymphoma (DLBCL), comparing them to normal SP controls. All malignancies analyzed had a lower mean expression of CAV1 compared to the healthy splenic tissue (Fig. 3c).

Fig. 3
figure 3

Follicle center B-cell lymphomas display altered CAV1 staining and expression patterns. a Representative tissue microarray cores stained for CAV1 expression. All images taken at ×20. b Bar graph of the distribution of composite scores assigned to healthy LN (n = 26), DBCL (n = 293), and FL (n = 18) samples for CAV1 staining. c Boxplot of the normalized log-transformed CAV1 expression value using the 212097_at signal probe for SP (n = 12), diffuse large B-cell lymphoma activated B cell-like (DLBCL-ABC) (n = 46), diffuse large B-cell lymphoma germinal center B cell-like (DLBCL-GCB) (n = 40), and FL (n = 27). Line within box represents median, whiskers extend to Tukey lines. (For (b): NS not significant, please see “Methods” for description of scoring key, *p < 0.05 Kruskal–Wallis and Dunn’s multiple comparisons test. For (c): NS not significant, *p < 0.001 comparison among groups of arrays method)

Full size image

Discussion

Herein, we include a characterization of Cav1 in the lymphoid compartments of a murine model. Cav1-deficient mice exhibited persistent splenomegaly compared to Cav1-heterozygous and WT mice. While the mechanism for Cav1-mediated splenomegaly is unknown, we ruled out a higher proliferative index as there were no differences found in Ki67+ staining numbers between genotypes. Previous investigations found very modest or no persistent differences in proliferation/apoptosis in Cav1-deficient spleens [5, 9, 14], suggesting a currently unresolved physiological condition may be causing the phenotype. However, a full knockout of Cav1 is required to drive the splenomegaly as our data suggests that Cav1-heterozygosity is not sufficient to drive the dysregulation as splenomegaly was not observed in Cav1+/− mice.

Previous investigations have established CAV1 as vital to T-cell function in the context of both antigen presentation and signal transduction [8, 11,12,13]. However, we and others demonstrate that Cav1 is seemingly dispensable for T-cell development and homeostasis [10, 11, 13, 14].

We present data in support of a role for Cav1 in the regulation of B-cell processes by observing a decrease in B-cell percentages, specifically in the SP, in aged Cav1-transgenic mice on the C57BL/6 background utilizing the knockout method described by Razani et al. [16]. This decrease was seen in conjunction with aberrant secondary follicle characteristics in aged mice, however a more detailed analysis of these changes is still needed. These findings are highlighted with our observation of a concurrent decrease in splenic B-cell percentages and an increase in inter-follicular plasma cells and fibrin deposits in Cav1-deficient and heterozygous mice without both exhibiting splenomegaly. This suggests that Cav1-heterozygosity is sufficient to drive the dysregulation of B-cell homeostasis and impair the Cav1-mediated internalization of fibrin [17], most likely caused by the angiogenic defects observed in Cav1-transgenic mice [16]. Further, the observation of increased inter-follicular plasma cells may elude to a basis for the decrease in CD19+CD22+ (i.e. mature, non-terminally differentiated) B cells in Cav1-transgenic mice without additionally perturbed populations as plasma cells (mature, terminally differentiated) downregulate pan-B surface markers during differentiation [18, 19]. Collectively, the decrease of mature B cells in conjunction with an increase of fibrin deposits and presence of plasma cells indicate that dysregulation of Cav1 may induce an inflammatory splenic microenvironment.

While there is controversy precise role of Cav1 in B-cell signaling, one point of agreement between all published reports is that there exists a role for Cav1 in B-cell processes as evident by conserved dysregulated responses and changes to physiological populations over time [14, 15]. However, the lack of consistency in the results derived from established Cav1-knockout models calls into question how the method and selection of exon as well as the genetic background of the mouse contributes to the observed results.

Taken together, the data suggests that loss of Cav1 alone is not sufficient to drive a robust immunological phenotype despite its frequent presence as a dysregulated molecule in human malignancies. However, Cav1-dysregulated cells appear “predisposed” to display aberrant signaling profiles given proper stimuli [10, 11, 13,14,15, 20,21,22]. This predisposition, or reprogramming, suggests CAV1 as a facilitator of immune/inflammatory-related processes and could point toward its ultimate role in the progression of multiple human malignancies, including both solid tumor [23,24,25] and hematological diseases [8, 26, 27]. To this end, we examined the staining pattern of CAV1 and its expression in follicle-derived B-cell malignancies. We observed fainter, less stromal-localized CAV1 staining in the hematological malignancies with lower CAV1 expression in tumor biopsies compared to normal splenic tissue. However, no differences were observed between the more aggressive DLBCL cases when compared to the more indolent FL cases, not forming a parallel between lowered CAV1 expression being associated with a more aggressive disease in both breast [23] and prostate cancer [24]. These results suggest that CAV1 is dysregulated in follicle-derived B cell malignancies but does not play an overt role between more aggressive and indolent disease entities.

Methods

Flow cytometry

Leukocyte populations were isolated using Lympho Separation Medium (MP Biomedicals, USA) with red blood cells lysed using RBC lysis buffer (Alfa Aesar, USA). Cells were counted and checked for viability (> 95%) using Trypan Blue (STEMCELL Technologies, CA) and 106 live cells were aliquoted for antibody staining (Additional file 1). Samples were run on a Cytoflex flow cytometer (Beckman Coulter, USA) and analyzed using the CytExpert 2.0 software (Beckman Coulter). Compensation was calculated utilizing a VersaComp antibody capture bead kit (Beckman Coulter). Cells were gated based off forward-side scatter, unstained controls, and fluorescence minus one controls. All samples are representative of at least 5 × 104 events.

Histology

Animal care protocol found in Additional file 6: Additional methods. Harvested organs were placed into 10% neutral buffered formalin and allowed to fix for ≥ 72 h before preservation in 100% ethanol. Preserved tissues were paraffin-embedded and sectioned (4 μm). H&E and Movat’s pentachrome staining were conducted using standard procedures, Ki67 staining using an auto-staining system.

Histological scoring was conducted for Ki67 staining, secondary follicle counting, and plasma cell identification. In each instance, three independent researchers scored the relevant variable in 4 separate fields of view (n = 3 per genotype). All slides were viewed at identical magnifications. For scoring agreement metrics and follicle area quantification, please see Additional file 5: Table S1, Additional file 6: Additional methods.

TMA scoring

The LY6161 high-density lymphoma and normal lymph node tissue array was utilized as previously described [27]. Cases were scored for both intensity of stain (faint, moderate, heavy) and type of staining observed (vessel only, stromal + vessel, or lymphocyte + stromal + vessel). Patient characteristics and scoring agreement described in Additional file 5: Tables S1, S2.

Gene expression data

(.CEL) files were downloaded via NCBI Gene Expression Omnibus and uploaded into BRB-ArrayTools. Samples were collated and normalized using the MAS5.0 method. Samples described in Additional file 5: Table S3.

Limitations

  • Unresolved splenomegaly in Cav1−/− mice.

  • More in-depth germinal center investigation requires additional IHC markers.

  • Higher resolution of B/T-populations could be achieved with additional flow markers.