Bavituximab Decreases Immunosuppressive Myeloid-Derived Suppressor Cells in Newly Diagnosed Glioblastoma Patients

This article is featured in Highlights of This Issue, p. 2949

Median overall survival (OS) in glioblastoma (GBM) is <15 months with standard chemoradiotherapy and adjuvant temozolomide (TMZ; ref. 1). Phosphatidylserine (PS) is a highly immunosuppressive membrane phospholipid on the inner leaflet of the plasma membrane in normal cells. In GBM cells and cells of the tumor microenvironment (including vascular endothelial cells), PS is externalized to the outer leaflet where it is recognized and bound by PS receptors on immune cells (2, 3). This leads to expansion of myeloid-derived suppressor cells (MDSC) and M2-like tumor-associated macrophages (TAM) that promote an immunosuppressive and pro-angiogenic tumor microenvironment (4).

Targeting PS in GBM xenograft models results in tumor cell death (5, 6) and has anti-angiogenic effects (7, 8), thus making it an attractive treatment strategy in GBM where neovascularization is prominent within the tumor microenvironment (9). Furthermore, PS-targeting agents reprogram the tumor microenvironment to an immune-activated state, including decreased MDSCs, TGFβ, and IL10 levels; increased TNFα and IL12 levels; repolarization of TAMs from an M2 to M1 phenotype; and induction of potent adaptive antitumor T-cell immunity (10, 11). In preclinical models, a synergistic effect between PS-targeting agents, radiotherapy, and TMZ has been observed (12, 13).

Bavituximab is a genetically engineered immunoglobulin gamma 1 (IgG1) chimeric mAb that binds β2-glycoprotein-1 (β2-GP1) to form a complex of β2-GP1 with PS, thereby reversing the immunosuppressive effect of PS (14). Bavituximab treatment in vitro and in vivo resulted in reduced tumor growth and prolonged survival that was enhanced by co-administration of chemotherapy and radiotherapy (3).

In this open-label, single-arm, phase 2 trial (NCT03139916), we evaluated the efficacy of bavituximab given with standard chemoradiotherapy and adjuvant TMZ in newly diagnosed GBM (nGBM). We assessed target-specific impact with serial perfusion MRI and by measuring myeloid-associated gene transcription and myeloid cell infiltrates at various timepoints during treatment to better understand the mechanism of action of bavituximab in GBM.

Eligibility criteria included: new diagnosis of histologically confirmed GBM or GBM variant (e.g., gliosarcoma); confirmation of wild-type IDH status by IHC or sequencing; age ≥18 years; no prior immunotherapy, alkylating agents, or radiotherapy to the brain; Karnofsky Performance Status (KPS) ≥60; life expectancy >6 months; adequate bone marrow, renal, and liver function; ability to undergo MR imaging; and dexamethasone dose <4 mg daily (or equivalent of another corticosteroid) at treatment initiation. Patients with both non-measurable and measurable disease by RANO criteria at baseline were eligible for inclusion. Exclusion criteria included: receipt of other investigational agents; known allergic reactions to compounds of similar chemical or biologic composition to bavituximab; concurrent use of moderate and/or potent enzyme inducers or inhibitors affecting bavituximab metabolism; uncontrolled intercurrent medical or psychiatric illness and social factors limiting compliance with study requirements; current pregnancy; known history of HIV, hepatitis B and C; history of other active malignancy (excluding in situ tumors) in the past 3 years; and known bleeding disorder/coagulopathy. Representativeness of study participants is shown in Supplementary Table S1. The study was conducted in accordance with the U.S. Common Rule.

This open-label, single-arm, phase 2 study conducted at the Massachusetts General Hospital and Dana-Farber Cancer Institute (NCT03139916) was approved by the Institutional Review Board of the Dana-Farber/Harvard Cancer Center and conducted in accordance with institutional and federal guidelines for human investigations. All participants were informed of the investigational nature of the study and provided written informed consent before enrollment.

Treatment started at least three weeks after craniotomy or at least two weeks after brain biopsy. Patients received fractionated involved-field radiotherapy (60 Gy, delivered in up to 2 Gy fractions per day) and concurrent TMZ (75 mg/m² per day) over 6 weeks, followed by six cycles of monthly adjuvant TMZ (C1-C6; 150–200 mg/m² on days 1–5 of each cycle; Supplementary Fig. S1). Bavituximab was given intravenously at 3 mg/kg weekly. This dose was selected on the basis of the data from phase 1 single-agent and phase 2 combination-therapy studies in which combination therapy did not substantially increase the risks of side effects. Bavituximab was administered starting day 5 of radiotherapy, for at least 18 weeks (i.e., until the end of C2), with the option to continue if tolerated. Treatment duration was based on the known synergism between bavituximab, radiotherapy, and TMZ and the diminishing effects of radiotherapy over time. Participants were pre-medicated with intravenous hydrocortisone 250 mg and diphenhydramine 50 mg if needed. To determine whether concurrent steroid treatment (including hydrocortisone pre-medication) interfered with response, patients were grouped by steroid dose during treatment with bavituximab (group 1: low-dose or no steroids; group 2: high-dose steroids; Supplementary Data).

The primary endpoint was proportion of patients alive at 12 months (OS-12). Secondary endpoints included progression-free survival (PFS; defined as time from treatment start to disease progression or death), OS (defined as time from treatment start to death or last follow-up, whichever occurred first), radiographic response, and bavituximab toxicity.

Patients underwent combined structural and physiologic MR imaging at baseline (pretreatment), pre-C1 TMZ (4 weeks post-chemoradiotherapy), and pre-C3 TMZ if they remained on treatment (“research MRIs”; Supplementary Fig. S1; Supplementary Table S2). Research MRIs included T2SPACE, FLAIR, dynamic contrast enhanced (DCE), dynamic susceptibility contrast (DSC), diffusion tensor imaging (DTI), and pre- and post-contrast T1-weighted MPRAGE images. DSC imaging included gradient echo (GE; reflecting panvascular or both micro- and macrovascular perfusion) and spin echo (SE; reflecting microvascular perfusion) images. After the pre-C3 MRI, patients could continue participation in research MRIs or undergo routine clinical MRIs every two months. Clinical MRIs included high-resolution T1-weighted post-contrast and FLAIR sequences; DTI and DSC imaging were not routinely acquired and therefore not included in the analysis after pre-C3. Response was determined per iRANO criteria by the treating physician (15).

For research scans, structural images (FLAIR, pre-/post-contrast MPRAGE) and parameter maps from DTI, DSC, and DCE imaging were registered to T2SPACE images using the BRAINSFit module in 3D Slicer (16). A deep-learning algorithm was used to segment contrast-enhancing tumor on MPRAGE images (excluding necrotic and hemorrhagic regions) and abnormal FLAIR-hyperintense areas on FLAIR images to generate the contrast-enhancing (CE) and FLAIR tumor regions of interest (ROI), respectively (17). ROIs were reviewed and manually adjusted as needed (K.I. Ly or E.R. Gerstner). Median values for ADC, Ktrans (a measure of vascular permeability), panvascular and microvascular relative cerebral blood volume (rCBV), and relative cerebral blood flow (rCBF) were calculated from contrast-enhancing and FLAIR-hyperintense ROIs. Ktrans maps (reflecting vascular permeability) were generated from DCE imaging. The detailed protocols for acquisition and post-processing steps for research scans are found under Supplementary Data.

Peripheral blood was collected pre-treatment, during week 6 of chemoradiotherapy, first week post-chemoradiotherapy, and monthly for two months during adjuvant TMZ/bavituximab (Supplementary Data).

Tumor tissue was obtained at diagnosis and progression (if available) and processed per standard pathology protocol. Analysis was performed on formalin-fixed paraffin-embedded (FFPE) tissue.

RNA was extracted from peripheral blood mononuclear cell (PBMC) lysates and FFPE tumor tissue using the Qiagen RNeasy Mini Kit (Qiagen; 74134). Quality and abundance of isolated RNA was measured using the Aligent Bioanalyzer 2100 and NanoDrop 2000 before sequencing. The nCounter Myeloid Innate Immunity Panel (probing for 770 myeloid-associated genes) was used to analyze their expression in PBMCs and tumor tissue (NanoString Technologies; ref. 18). All raw data were reviewed for quality using the nCounter Advanced Analysis 2.0 software (NanoString Technologies). Sequencing results were uploaded to ROSALIND (https://rosalind.onramp.bio/) for analysis and plot visualization (ROSALIND Inc.). Normalization, fold changes, and P values were calculated on ROSALIND using criteria from the nCounter Advanced Analysis 2.0 protocol. Cutoff values for P value and fold change were 0.05 and +1.5/−1.5, respectively. We entered Nanostring-derived gene expression data into Metascape (19), an open-access resource (www.metascape.org) that provides expression analysis and defines over-represented gene sets, or “enrichment clusters.”

FFPE blocks with tumor cell percentages >80% were selected before sectioning. Slides were stained for intratumoral myeloid cellular infiltrate, focusing on the number of TAMs and MDSCs (Supplementary Data). Slides were scanned using the Vectra Polaris microscope and unmixed with inForm software (Akoya Biosciences), incorporating the algorithm generated by a spectral library using single-fluorophore label control slides. Unmixed images for each slide were fused and uploaded to the Halo software (Indica Labs) for analysis. The entire tumor tissue area was manually annotated, excluding folds or debris. Cell segmentation and staining intensity thresholds were assigned by a blinded individual. Parameters defining each cell type were applied to the analysis algorithm to each slide. M1 or M2 TAMs were identified on the basis of co-staining for CD68 and CD163 (CD68⁺CD163⁻: M1 TAMs, CD68⁺CD163⁺: M2 TAMs). Granulocytic MDSCs were characterized by co-staining with CD15 and Lox-1 (a differentiation marker for human MDSCs; refs. 20, 21). Positive cell counts were measured as number of positive cells per mm² of tissue.

All statistical analyses (except for the immune marker analysis) were performed using SAS9.4 (SAS Institute Inc.). OS-12 in recent studies of novel agents in nGBM was 60%–68% (22, 23). This study was designed to compare OS-12 against a historic control of 60% using the method of Brookmeyer and Crowley (24). A sample of 33 patients would provide 0.88 power (target = 0.85) to detect OS-12 of 80%, given a one-sided alpha of 0.09 (target = 0.1). The null hypothesis would be rejected if the observed Kaplan–Meier estimate of OS-12 was ≥72%. To allow for early patient drop-out or withdrawal, 36 patients were enrolled. Patients lost to follow-up before documented progression or death were censored at the time of their last visit.

The relationship of clinical (MGMT status, age, steroid group) and imaging markers with OS-12 was tested using univariate logistic regression. The correlation between markers and PFS/OS was tested using Cox proportional hazards model. For the multivariate analysis, three imaging markers (which were significant in univariate analyses and of clinical relevance) were included after excluding high correlation between markers (using Pearson correlation coefficient). To assess the association of changes in imaging markers with OS-12, PFS, and OS, we calculated the fold-change between pre-C1 and baseline (fold change = timepoint/baseline). Values were dichotomized into “low” versus “high” at the median of their distribution to ensure an unbiased division and equal number of patients in each group. No adjustment was made for multiple testing, given that these were exploratory and hypothesis-generating analyses.

For the immune marker analysis, GraphPad Prism 9.1 software was used for statistical analyses and visualization. We categorized patients into “low” and “high” PFS/OS groups, divided at the median. All tests were two-sided with a P value of <0.05 considered statistically significant. We used unpaired t and Mann–Whitney U tests to assess differences between timepoints.

The data generated in this study may be available upon request from the corresponding author.

Thirty-six patients with IDH-wt nGBM were enrolled between July 2017 and 2018 (Table 1). Of these, 33 patients were evaluable; 2 were not evaluable because of withdrawal of consent and 1 experienced clinical progression before start of treatment. Median age was 59 years and median KPS was 90. Fifty-eight percent of tumors were MGMT promoter-unmethylated. Fifteen patients (45.5%) did not require dexamethasone at treatment initiation and during treatment. Ten patients (30.3%) required low-dose or no steroids and 23 patients (69.7%) required high-dose steroids during treatment. Six (18.2%) patients had a biopsy, 12 (36.4%) patients underwent a subtotal resection, and 15 (45.5%) patients had a gross total or near–gross total resection (Supplementary Data).

The proportion of patients alive at 12 months (OS-12) was 72.7% [95% confidence interval (CI), 59%–90%; 24 patients alive at 12 months], achieving the study's primary endpoint. At the time of data cutoff (May 2021), median PFS and OS were 6.9 and 15.4 months, respectively (Supplementary Fig. S2). Thirty (90.1%) patients progressed. Of 3 patients without disease progression, one died of aspiration pneumonia, one is still on trial, and one has been lost to follow-up. Twenty-eight (84.8%) patients have died, 4 are alive, and one has been lost to follow-up. Twenty-six of 33 patients (78.8%) had measurable disease at baseline per RANO criteria. Stable disease was observed in 79%, partial response in 15%, and progressive disease in 9% of patients. In those with a partial response, the median duration of response was 177 days (range, 55–225 days; Supplementary Fig. S3). Methylated MGMT status was associated with improved PFS (P = 0.003) and OS (P = 0.002) but age and steroid use were not (Supplementary Table S3).

The median number of bavituximab doses given was 24 (range, 5–85). Five patients stopped after a median of 18 doses (range, 6–28) due to personal preference. One patient stopped after 14 doses due to repeated hypersensitivity reactions (facial flushing, nausea) despite standard pre-medication.

Twenty-eight (84.8%) patients had adverse events (AE) that were possibly, probably, or definitely related to bavituximab. AEs observed in at least 10% of patients are shown in Supplementary Table S4. Nine AEs in five patients resulted in interruption of treatment with bavituximab (Supplementary Table S5). In one patient with grade 2 fatigue thought to be possibly related to radiotherapy, treatment was permanently discontinued. Four grade 3 or 4 AEs and no grade 5 were reported.

On univariate analysis, high baseline CE volume, FLAIR volume, panvascular and microvascular rCBV, panvascular rCBF, and Ktrans in the CE region were associated with worse OS (Table 2). There was a trend toward significance for Ktrans and OS-12 (OR, 9.33; 95% CI, 0.96–90.94; P = 0.054). High baseline panvascular rCBV and Ktrans within the CE region were predictive of worse PFS. On multivariate analysis, we stratified patients by MGMT status and included CE volume and panvascular rCBV as predictors, given that these are known biologic markers of tumor aggressiveness and vascularity. CE volume remained associated with worse OS (P = 0.04). A trend was also observed between high panvascular rCBV and worse OS (P = 0.064). High panvascular rCBV remained a negative predictor of PFS (P = 0.047).

At the pre-C1 timepoint, the fold-change of imaging markers compared with baseline was analyzed (Fig. 1; Supplementary Table S6). The median fold-change in panvascular rCBF was 0.76. Patients with a fold-change ≥0.76 had a three times greater risk of death compared with those with a fold change <0.76 (P = 0.018; Table 2). For Ktrans, the median fold-change was 1.34. Patients with a fold-change ≥1.34 had an 83% reduction in the hazard of death compared with those with a fold change <1.34 (P = 0.003; Table 2). On multivariate analysis stratified by MGMT status, both markers remained significant: a pre-C1 increase or stable panvascular rCBF predicted worse OS whereas increased Ktrans was associated with a reduction in death (Supplementary Fig. S4). At the pre-C3 timepoint, the fold-change of imaging markers was not associated with OS-12, OS, or PFS on univariate analysis.

We performed gene expression profiling and analysis of tumor specimens and PBMCs. At baseline (pre-treatment), RNA analysis of tumor specimens (n = 10) showed significantly positive shifts in myeloid-related gene expression in patients with longer survival, with enrichment of 116 and 120 transcripts and downregulation of 1 and 2 gene(s), in patients with longer PFS and OS, respectively (Fig. 2A–C). Using Metascape cluster enrichment analysis of overexpressed genes, the most significantly upregulated networks in longer surviving patients were associated with inflammatory response, cytokine signaling, and leukocyte activation (Fig. 2B–D).

In comparison, PBMC gene expression data from 26 patients at baseline showed little differential transcription of myeloid-related genes based on survival. Eleven and 18 genes were differentially expressed in high versus low PFS and OS groups, respectively. Approximately equal numbers were upregulated or downregulated (Supplementary Fig. S5).

We also compared RNA transcripts from tumor specimens at diagnosis (n = 10) and at progression (n = 7), including paired samples (pre- and post-treatment) from 2 patients. Sixty-seven myeloid-related genes were differentially expressed between these two timepoints: 35 and 32 genes were upregulated and downregulated, respectively, at progression (Fig. 3A). Metascape cluster enrichment analysis of differentially expressed genes (Fig. 3B) revealed that the most commonly upregulated and downregulated functional networks included shifts in expression of IL4, IL10, and IL13, all of which are associated with MDSC maturation and signaling (25).

We performed multispectral immunofluorescence to characterize the phenotype of TAMs in tumor tissue at diagnosis (n = 10) or progression (n = 8). There was no difference in the density of TAMs or M2/M1 ratio in tumor tissue between these timepoints (Supplementary Fig. S6). However, we found decreased MDSCs in tumor tissue after bavituximab treatment (4.5+2.3 cells/mm²) compared with time of diagnosis (63.1+4.5 cells/mm²; Fig. 4; Supplementary Fig. S7).

We evaluated the efficacy of bavituximab with standard chemoradiotherapy and adjuvant TMZ in 33 patients with IDH–wild-type nGBM. The study met its primary endpoint with an OS-12 of 72.7%. However, median PFS (6.9 months) and OS (15.4 months) were comparable with standard treatment in combined methylated and unmethylated patients with GBM (1). Bavituximab was well tolerated and not associated with significant AEs. Among current clinical reports of bavituximab in cancer (26–29), our study is the first to directly and prospectively probe tumor microenvironment and assess on-target effects of bavituximab using perfusion MRI and expression of immune-related genes and relevant immune cell subsets.

Baseline elevation and interval increase or stable panvascular rCBF were associated with decreased survival, consistent with previous studies in nGBM (30, 31). Whereas these studies found that a pre-C1 decrease in vascular permeability (measured by Ktrans) was associated with improved survival, we found that an increase in Ktrans was associated with improved outcome. This may suggest a greater immune response causing breakdown of the blood–brain barrier and, consequently, better penetration of bavituximab to achieve on-target effects within the tumor microenvironment.

The GBM immune microenvironment is characterized by relatively few CD8⁺ T cells and abundant macrophages (32). M2 TAMs, the predominant cells in the GBM microenvironment, promote immune tolerance (33). Cellular interactions in this M2-skewed microenvironment enhance MDSC differentiation that further inhibit antigen-specific T-cell responses. In animal models, bavituximab-mediated PS blockade leads to reprogramming of TAMs from an immunosuppressive M2 phenotype to an immune-active M1 phenotype and reduces MDSCs (3, 14), with the latter phenomenon confirmed in patients with GBM in this study. Furthermore, our finding that myeloid transcripts were more highly expressed in tumors from patients with better survival supports that bavituximab might have greater activity when the drug's targets are enriched. In GBM subsets, increased macrophage and microglia infiltration has been associated with worse survival (34). As expected, expression levels of myeloid-related genes in PBMCs did not predict survival as bavituximab exerts its impact within the tumor microenvironment and not in PBMCs. Accordingly, we found that MDSC number in tumor specimens was reduced at time of progression, confirming an on-target, intratumoral effect of bavituximab and a potential step forward in the modulation of the GBM immune microenvironment.

Our study has some limitations. The single-arm design does not permit evaluation of the differential effects of bavituximab and standard therapy on immune cell modulation and 6/30 patients discontinued treatment, although this was less than with other bavituximab studies. Tumor specimens were obtained from only 21% of patients at progression so may not be representative of the entire cohort or accurately reflect the tumor microenvironment during treatment. However, GBM progression is typically associated with at least maintained, if not greater, local immunosuppression and increased numbers of M2-TAMs and MDSCs (35). Therefore, our finding of decreased MDSCs at progression likely reflects on-target treatment effects, and tumor progression presumably occurred because of molecular and/or immune escape. Bavituximab typically suppresses IL10 expression so our finding that IL10 was upregulated at progression suggests an immune escape process. A “window of opportunity” trial focusing on changes in TAMs, MDSCs, and lymphocyte subsets during bavituximab treatment might clarify peak immune changes associated with this drug (32, 36). Our target OS-12 was generated from historical controls and was a conservative target because we enrolled both MGMT methylated and unmethylated patients that are now known to have distinct prognoses. However, for this first-in-GBM study, we believed this was a reasonable initial target endpoint.

In summary, this mechanistically driven phase 2 trial is the first to demonstrate on-target effects of bavituximab with a decrease in intratumoral MDSCs. Greater expression of myeloid transcripts in pre-treatment tumor tissue correlated with PFS and OS. Further study is warranted to identify those most likely to benefit from bavituximab and optimal combination therapies. Higher serum β2-GP1 levels have been explored as a marker of response in patients with non-squamous non–small cell lung cancer (NSCLC; ref. 26) but further validation of β2-GP1 levels as a predictive biomarker will be required. In addition, in a phase 3 trial of non-squamous NSCLC, post hoc analyses revealed that OS was improved in a small subset of patients who received immune checkpoint inhibition after docetaxel and bavituximab (26). As a result, several trials of bavituximab and pembrolizumab in solid cancers are ongoing (NCT04099641, NCT03519997, and NCT04150900). In GBM, pembrolizumab monotherapy or combined with bevacizumab has shown limited benefit. However, on the basis of the above data, a potential next step in GBM is to combine bavituximab with checkpoint inhibition or other immune-activating approaches such as vaccination or chimeric antigen receptor T-cell therapy to maximize the intratumoral immune response.

D.A. Reardon reports personal fees from Agios, AnHeart Therapeutics, Avita Biomedical, Inc., Bristol Myers Squibb, Boston Biomedical, CureVac AG, Del Mar Pharma, DNAtrix, Hoffman-LaRoche, Ltd., Imvax, Janssen, Kiyatec, Medicenna Therapeutics, Neuvogen, Novartis, Novocure, Pyramid Biosciences, Sumitomo Dainippon Pharma, Vivacitas Oncology, Y-Mab Therapeutics, and Blue Rock Therapeutics outside the submitted work. I.C. Arrillaga-Romany reports grants from NCCN during the conduct of the study; I.C. Arrillaga-Romany also reports personal fees from Boehringer Ingelheim, as well as other support from Astex Pharmaceutical and GSK outside the submitted work. D.A. Forst reports other support from Eli Lilly outside the submitted work. J.T. Jordan reports personal fees from Navio Theragnostics, Alexion Pharmaceuticals, Recursion Pharmaceuticals, Shepherd Therapeutics, and CEC Oncology outside the submitted work. J. Dietrich reports personal fees from Unum Therapeutics and Amgen outside the submitted work. L. Nayak reports grants from Leukemia and Lymphoma Society; personal fees from Ono, BraveBio, and Genmab; and non-financial support and other support from Merck, AstraZeneca, Kazia, and Bristol Myers Squibb outside the submitted work. T.T. Batchelor reports grants from National Institutes of Health outside the submitted work. J. Kalpathy-Cramer reports grants from GE Heathcare and Genentech, as well as other support from Siloam Vision Inc. outside the submitted work. E.R. Gerstner reports grants from NCCN during the conduct of the study. No disclosures were reported by the other authors.

K.I. Ly: Data curation, formal analysis, investigation, visualization, writing–original draft, writing–review and editing. L.G. Richardson: Formal analysis, investigation, visualization, writing–original draft, writing–review and editing. M. Liu: Formal analysis, writing–original draft, writing–review and editing. A. Muzikansky: Conceptualization, formal analysis, writing–review and editing. J. Cardona: Investigation. K. Lou: Formal analysis, investigation. A.L. Beers: Investigation. K. Chang: Investigation. J.M. Brown: Investigation. X. Ma: Investigation. D.A. Reardon: Investigation, writing–review and editing. I.C. Arrillaga-Romany: Investigation, writing–review and editing. D.A. Forst: Investigation, writing–review and editing. J.T. Jordan: Investigation, writing–review and editing. E.Q. Lee: Investigation, writing–review and editing. J. Dietrich: Investigation, writing–review and editing. L. Nayak: Investigation, writing–review and editing. P.Y. Wen: Investigation, writing–review and editing. U. Chukwueke: Investigation, writing–review and editing. A. Giobbie-Hurder: Formal analysis, writing–review and editing. B.D. Choi: Investigation. T.T. Batchelor: Investigation. J. Kalpathy-Cramer: Investigation. W.T. Curry: Conceptualization, investigation, writing–original draft, writing–review and editing. E.R. Gerstner: Conceptualization, funding acquisition, investigation, visualization, writing–original draft, project administration, writing–review and editing.

This study was funded by the National Comprehensive Cancer Network Oncology Research Program and from general research support from Oncologie, Inc. (E.R. Gerstner). In addition, funding was provided by A Shot for Life (https://www.ashotforlife.org/) and The Jenny Fund (https://www.thejennyfund.org/; to W.T. Curry).

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.