Quantification of bone marrow interstitial pH and calcium concentration by intravital ratiometric imaging

The fate of hematopoietic stem cells (HSCs) can be directed by microenvironmental factors including extracellular calcium ion concentration ([Ca2+]e), but the local [Ca2+]e around individual HSCs in vivo remains unknown. Here we develop intravital ratiometric analyses to quantify the absolute pH and [Ca2+]e in the mouse calvarial bone marrow, taking into account the pH sensitivity of the calcium probe and the wavelength-dependent optical loss through bone. Unexpectedly, the mean [Ca2+]e in the bone marrow (1.0 ± 0.54 mM) is not significantly different from the blood serum, but the HSCs are found in locations with elevated local [Ca2+]e (1.5 ± 0.57 mM). With aging, a significant increase in [Ca2+]e is found in M-type cavities that exclusively support clonal expansion of activated HSCs. This work thus establishes a tool to investigate [Ca2+]e and pH in the HSC niche with high spatial resolution and can be broadly applied to other tissue types.

out several issues regarding accuracy and reliability of the method, as discussed in the following.
The authors, from a lab having pioneered in vivo two-and three-photon microscopy of calvarial bone and bone marrow, chose for their calcium imaging approach Rhod-5N due to its high Kd, necessary to quantify interstitial calcium, also adjecent to bone. They coupled this red fluorescence calcium dye with the green emitting Alexa 488 dye to enable ratiometric measurements and correct the spectral signal ratios for pH value and effects of Rayleigh scattering of bone and marrow separately and developed an easy to use approach which I expect to find broad application in the bioscientific/biomedical community. The descrinbed normalization steps of the spectral signals is absolutely necessary, however, in order to ensure the reliability of the determined absolute calcium concentrations retrived by the presented method further validation is needed. Especially, validation using a different fluorescence method -fluorescence lifetime imaging -, which can be applied in vivo and is, in general, less affected by experimental circumstances would make the approach much stronger. While certainly not perfectly fitting, here are some suggestions: fluorescence lifetime imaging of CaG5N has been succesfully employed in measuring interstitial calcium levels in skin ( 1. A full titration curve of the pair Rhod-5N / Alexa 488 with not only the characteristic Kd but also Hill-slope is needed in order to characterize the sensitivity of the approach in different regions of the calcium dynamic range -at the edges (asymptotic parts of the curve, the reliability of the results is much lower). Statistics (e.g. Man-Whitney-test) weighs all calcium concentrations disregarding the shape of the titration curve. Eventually, this would clarify the lack of heterogeneity of calcium levels throughout the different LT-HSC niches.
2. Also for SNARF-1 such a full titration curve depicting the sensitivity throughout the pH dynamic range would strengthen the method.
3. Except for one example, interstitial/extracellular calcium maps at only one single time-point have been acquired and used in the correction algorithm taking into account the different (mainly) Rayleigh scattering of the red (Rhod-5N) vs. green (Alexa 488) emitted fluorescence. However, heterogeneity will certainly occur also over time -this aspect needs to be taken into account when validating the correction algorithm. I expect that effects of different photobleaching behaviourand, in general, photophysical behaviour -of the two dyes will change their spectral signal ratio and, thus, the determined calcium concentrations (as well as the pH values determined by SNARF1). Especially, as published by the lab, the oxygen levels (pO2) strongly vary throughout the bone marrow -this has an impact on the cells, but also some order of magnitude below in scale, on the dye molecules too: the fluorescence deplition as well as spectral shifts of excitation and fluorescence spectra (just as an example, Stockes / anti-Stockes shifts) are dramatically influenced by the depletion of the first triplet state of the dyes -how much this is populated due to inter-system crossing in Rhod-5N vs. Alexa 488 needs to be taken into account. Hence, the influence of local oxygen concentrations on the fluorescence signals ratio (and on calcium levels) needs to be verified for full accuracy. Of course, the same hold true for SNARF-1, since differently ionized forms of the same molecule are expected to have different energetical levels (not only ground and first singlet state, but also different triplet states, imposing a change in the intersystem crossing rate and possible effect of oxygen). 4. The authors correct using a bi-exponential function with tissue depth (z) for the high-frequency scattering through bone and bone marrow, however, for me it was unclear how the already published mean free-scattering paths (or EAL -e.g. as published by the Chris Xu lab, Ozounov et al, Nat. Meth. 2018; Wang et al, Nat. Meth. 2019) needed to describe this type of scattering are included since the constants C1 and C2 are not futher described in the text. 5. Additionally, both scattering effects and effects of wave front distorsions (lens effects of blood vessels, spherical aberrations, astigmatism) are not only depth-dependent but vary also within single tissue layers (in x and y). Whereas a quantification would be extremly tidious, at least estimating the impact of these effects would have on the accuracy of pH values and calcium levels is necessary. The accuracy will be impaired -the question is, if relevant differences between different regions in the bone marrow can still be detected, given the uncertainty caused by these effects. 6. Last but not least, the survival niches of HSCs (but also of other immune cells) suffer changes not only on the short time-scale (minutes to hours) but especially on the longer time-scale, as shown also by the Lin lab in longitudinal imaging experiments of the calvarial marrow. Since, among others, changes in pO2 are expected due to a continuos change in the position of (various types of) blood vessels in the marrow (especially in long bones (Reismann et al, Nat. Comm. 2017) but also in the calvarial bone), a comment of how the method will deal with such changes for a full characterization of the extracellular calcium levels within survival niches is required.
Reviewer #3: Remarks to the Author: In this manuscript Yeh and colleagues report the development of an intravital imaging-based technique that allows to perform spatially resolved measurements of pH and calcium in the bone marrow (BM) microenvironment. The approach makes use of calcium sensitive ratiometric probes, and previously established, widely used multiphoton intravital imaging of calvarial BM tissue in mice, in which the authors are experts. Notably, the method requires the implementation of mathematical correction of the attenuation of fluorescence with imaging depth of the probes, which depends on the both the thickness of the bone as well as that of marrow tissue that light needs to go through. Using this technique, the authors are able to provide measurements of the interstitial pH and [Ca2+] in BM tissues and they report differences between endosteal regions depending on the metabolic state of the proximal bone surface. Finally, by using a reporter mouse of hematopoietic stem cells they also assess the pH and [Ca2+] in the immediate vicinity of HSCs, thus providing estimation of these parameters in the HSC niche . The BM is the primary site for hematopoiesis and hematopoietic stem cell maintenance. A critical question, which remains unresolved to date, is the specific cellular and molecular composition of the anatomical niches in which HSCs reside. Furthermore, to what extent different spatial compartments of the marrow differ in their physiological conditions is of great interest to the understanding of spatial compartmentalization in this tissue. Thus, the ability to simultaneously perform spatially resolved measurements of key parameters dictating cell fate, such as oxygen levels, [Ca2+] or pH, in situ and in vivo and in a non-invasive fashion, is of great relevance, technically very challenging and of high merit. I have some comments on specific points, which in my view could improve the manuscript • The main caveat to the study is acknowledged by the authors in the Discussion. The need to pair the calcium indicator probe with a reference dye requires that both dyes have the same biodistribution, or otherwise this could lead to inaccurate measurements depending on the tissue region imaged. While it could be assumed that both dyes used in the study may indeed not vary much in their biodistribution, this is difficult to ascertain at this point and casts doubt as to the accuracy of the data. Indeed, from the images in Supplemental video 3, it would seem as the dyes are not always evenly distributed. In the discussion, the authors propose an elegant way to circumvent this, the coupling of probes to low molecular weight dextrans, whose biodistributions would be equal. Given that dextran conjugation is relatively straightforward, why is this approach not tested here to ultimately confirm the validity of their technique? When possible it would be desirable to have this experiment done and the measurements repeated and compared to the dextran-free approach. • In the introduction the authors mention that local concentrations of [Ca2+] in distinct regions of the BM (endosteal) can reach really high levels, pointing to the existence of substantial differences between different tissue compartments. According to Figure 3c, [Ca2+] in the interstitium does vary over a wide range and therefore it would be important to understand whether these variations are related to specific localization with regards to relevant anatomical landmarks such as bone surfaces. While in the discussion it is mentioned that a gradient of [Ca2+] towards endosteal surfaces is not detected this is not clearly shown in the figures. It would be important to depict the values of [Ca2+] as a function of distance to bone, and similarly address how [Ca2+] vary with the distance to different blood vessels (sinusoids/arteries), which have been proposed to harbor distinct niches for specific hematopoietic populations.
• Similarly, are there local differences in pH between different regions of the BM? According to Figure 1f the pH values in the interstitium range from 6.8 to 7.4. Are these variations related to spatial location? • In the discussion the authors mention that LT-HSCs were not found in areas with lowest [Ca2+] levels, however this trend is not quantified and shown in the Figure. • Along the same lines , the measurements of both pH and [Ca2+] in the vicinity of HSCs are interesting, but do not inform on whether the values for both parameters are exclusive or distinct for HSCs, or all cell types in the BM are exposed to similar conditions. The authors could use reporter mice for other cell types, for instance, B cells, T cells or neutrophils and perform similar measurements that can be used as a reference to understand this issue.
• The study would gain in significance if the authors assessed how conditions in which hematopoiesis is drastically altered, modify the interstitial values of pH and [Ca2+]. For instance, how does treatment with 5-FU, a myeloablative drug alter these parameters in the BM Minor points: • The authors should provide details on the spatial resolution in all dimensions of these measurements of pH and [Ca2+] in the main text. • In Figure 2a, it would be interesting to provide the SHG image to visualize the collagen signal of the areas marked as osteoids. • In Figure 2c the legend of the x axis is duplicated We thank the reviewers and the editor for their helpful comments on our manuscript, "Quantification of bone marrow interstitial pH and calcium concentration by intravital ratiometric imaging". To address the concerns raised by the reviewers, we have performed additional experiments and provided a substantial amount of new data, which we believe have significantly improved the manuscipt. Please see below for the point-by-point response.

Reviewer #1
The manuscript by Yeh et al. described imaging approach to perform absolute quantification of pH and calcium concentration in mouse calvarial bone marrow. This study has some significance from the perspective of advanced technique that allows us to evaluate dynamic pH and Ca2+ concentration in living mice. However, the authors did not demonstrate how low or high Ca2+ concentration affects HSC biology, although they just described Ca2+ concentration as a microenvironmental factor for HSCs throughout the manuscript. The relationship between Ca2+ concentration and the HSC localization has left unclear. There are some specific comments.
Response: We thank the reviewer for the comment on the technical advance described in this work, which we believe will be broadly useful beyond measuring pH and [Ca 2+ ]e, as an increasing number of ratiometric sensors are being developed for functional imaging and probing the tissue microenvironment in vivo. Clin Invest. 131(4), 2021] have attempted to evaluate the mTOR activity and extracellular ATP levels in the bone marrow of live mice. While very powerful, proper use of these sensors require rigorous analysis of their fluorescence signals as they propagate through tissue, a fact that has not been recognized in the field so far. We believe it is important to bring to attention the need for analytical methods that can recover the correct ratios in order for these measurements to be meaningful and quantitative. We have added this to paragraph 1 of the Discussion section. ]e (as described in this manuscript) but also to isolate them from these specific locations under image guidance for molecular profiling (e.g. single cell RNA-seq) and functional assays. That technology is currently under development in our laboratory but is beyond the scope of this paper.

Major comments:
1) In order to perform depth correction, the authors implemented a two-step algorithm. However, the validity and reliability of the method remains to be insufficiently clear. They should use a more robust parameter such as fluorescence lifetime.

Response:
We agree that it is critical to demonstrate the validity and reliability of the two-step algorithm for depth correction. To this end we have performed additional validation experiments using a dye (Rhodamine-B dextran (70kDa)) that is independent of calcium and pH within the physiologic range. We split the relatively broad emission spectrum of this single emitter into two (green and red) channels, reasoning that the red-to-green (R/G) ratio should be invariant with depth and location in tissue. We first verified that the in vitro R/G ratio of the Rhodamine-B dextran is independent of laser power, dye concentration, [Ca 2+ ], or pH from 6.6 to 7.3 (Fig R1 a). We then performed in vivo imaging of bone marrow labeled with Rhodamine-B dextran and observed a similar red shift with increasing imaging depth as observed with Snarf-1 and the AF488/Rhod-5N dye pair. Using the same depth correction algorithm as described in the manuscript, we were able to correct the red shift and recover the intrinsic R/G ratio, free from tissue optics-induced spectral distortion. The recovered value is in agreement with the R/G ratio of the Rhodamine-B dextran measured in vitro (Fig R1 b- We have performed an additional validation experiment by thinning the bone with femtosecond laser-mediated ablation (Fig R1 e-f). We acquired 3D stacks of the same Rhodamine-B dextranlabeled bone marrow cavity (same field of view) just before and just after laser bone thinning. As shown in panel Fig R1 g-h below, the same dye exhibited different red shift depending on the thickness of the bone above it, attesting to the tissue-optics origin of the red shift. Importantly, applying the two-step correction, we are able to recover the same R/G ratio as the ratio determined for Rhodamine-B in vitro. The recovered intravascular R/G ratio (mean = 0.59 ± 0.034) is consistent with the in vitro measurements (ratio = 0.59 ± 0.005). No significant difference was observed between vessels and interstitium (ratio = 0.58 ± 0.021) (n = 30 and 31 sub-ROIs from vessels and interstitium, respectively. N = 1 mouse). Two-sided Mann-Whitney test. Mean ± s.d. (e)-(f) Cross-sectional images of calvarial bone and bone marrow before and after laser bone thinning. (g) Intact bone and thinned bone exhibited a different extent of red shift and were recovered after the two-step depth correction. (h) The absolute R/G ratios plotted as a function of depth before correction, after bone thickness correction, or after completing the twostep depth correction. After bone thickness correction, the ratios at the endosteum from both intact and thinned were recovered close to the measured values in vitro. After both thickness and depth corrections, the ratios at all depths were recovered and consistent with the in vitro measurements.
While we agree that FLIM is generally considered a more robust method than ratiometric imaging, we do not think it is suitable for our purpose because of multiple reasons: 2. To our knowledge, Rhod-5N (Kd ~320 µM) is the only available calcium sensor that responds in the mM range, and we are not aware of any report of Rhod-5N fluorescence lifetime measurement in the literature. To be a useful FLIM-based calcium sensor, the dye needs to have reasonable fluorescence quantum yields in both its calcium-free and calcium-bound forms. Since the quantum yield of one of them (the calcium-free Rhod-5N) is very low, it will be extremely unlikely that its fluorescence lifetime can be used for accurate bi/multi-exponential fitting or phasor analysis.
3. Even if we were able to find a suitable dye for FLIM, getting sufficient photon count is always a challenge, especially for in vivo imaging. Good signal to noise ratio is essential for obtaining accurate determination of fluorescence lifetimes. The reported integration times for FLIM imaging of the skin (a much less scattering tissue than the bone) is 2-3 min per x-y frame (256x256) [Celli et al. Biophysical Journal, 98, 911-921, 2010], or 60-90 min for a single z stack, making it impractical for in vivo 3D imaging given the rapid clearance of the dyes from the tissue. By comparison, our acquisition time for ratiometric imaging is 2 sec per frame (512x512) and ~1 min per z stack. 4. It is not clear whether FLIM is really a more robust method. Fluorescence lifetime itself can be sensitive to oxygen, polarity, pH, etc. In addition, because of the long integration time required, it is more prone to photobleaching, triplet state built-up (ground state depletion), and subject to dye clearance.
We have added this in a new paragraph 9 of the Discussion.
2) To convert the ratios measured in vivo such as R/G and Rhod5N/AF488 to pH and calcium concentration, respectively, the authors established a calibration curve showing the relationship between them in vitro. To make the clear the validity of the conversion, they should provide the evidence that the correlation between the ratios and pH/Ca completely matched in both in vitro and in vivo.

Response:
We thank the reviewer for the comment. We understand that an in vivo calibration curve will be the most ideal, however it is not feasible because the physiologic pH or [Ca 2+ ] are set within a narrow range and the animal will exhibit detrimental cardiovascular responses if attempting to titrate the serum levels out of the physiological range. We can only show that the intravascular pH and [Ca 2+ ] obtained by our in vivo ratiometric measurement are in agreement with the values obtained by ex vivo electrode measurements of the extracted blood serum, but the calibration curves over the full range of pH or [Ca 2+ ] have to be established in vitro. As the depth correction algorithm was able to recover the intrinsic ratios in the serum that are in agreement with both in vitro measurements (Fig. R1) and literature values, we believe these results support the validity of the in vitro to in vivo conversion.
We wish to point out that in our original manuscript, the calcium levels were reported using the calibration curve established with the Arsenazo assay following the convention used by Fig 3b in the manuscript), which measures the total calcium concentration in the medium. However, after consulting with experts in the field, we now think it is more appropriate to report the free calcium ion concentration because the calcium buffering capacity can vary and is unknown. Therefore, we have established a new calibration curve using [Ca 2+ ] measured by an ion-selective electrode (new Fig 3b in the revised manuscript). All data reported in the revised manuscript is based on the new calibration curve.
3) The authors employed Rhod-5N to visualize extracellular calcium in vivo and used AF488 as a reference dye. Do Rhod-5N and AF488 have similar ADME characteristics? If not, AF488 is not suitable as a reference dye.

Response:
We agree with the reviewer's concern that any differences in the ADME (absorption, distribution, metabolism and excretion) characteristics of Rhod5N and AF488 would change the measured ratios significantly. As stated in the original manuscript and also pointed out by Reviewer 3, this concern can be addressed by conjugating both Rhod-5N and AF488 to a dextran backbone so the ADME would be identical for the dye pair. However, we were advised by our colleagues in chemistry that conjugation of Rhod-5N to dextran would alter its calcium response, as the same moiety for conjugation is also used for calcium binding. Instead, we packaged Rhod5N and AF488 with a fix stoichiometry into micelles (~125 nm hydrodynamic radius), so the packaged dyes had the same biodistribution, and were cleared at the same rate. We validated the ratiometric measurement using a strict comparison: the animal was injected with micelles first for obtaining R/G ratios. Once the micelles were cleared (~4 hours), we injected our original dye mixture (with two separate dyes) to obtain the R/G ratio of the same bone marrow cavity for comparison. In Fig. R2 below, we show the close agreement between the co-injected Rhod5N and AF488 and the micelle packaged dye mixtures, indicating that Rhod5N and AF488 have similar biodistribution in vivo, as verified by the consistent R/G ratios from the same ROIs. We have included the results as a new Suppl. Fig. 7, as well as the methodology and characterizations of micelles in the revised manuscript. 4) Rhod-5N, calcein blue and tetracycline bind to calcium. Can Rhod-5N be used in combination with calcein blue and tetracycline? Is there any possibility that the sensitivity of Rhod-5N to calcium attenuated by them?
Response: Thank you for raising this point. Calcein blue was injected 2 days before imaging. Its effect was transient (Fig 2d in the manuscript) and did not alter the serum calcium on the day of imaging as verified by the Arsenazo assay. Alizaren red was injected on the day of imaging but after the acquisition of the Rhod5N/AF488 data so it does not interfere with the calcium measurement. This is now more clearly stated in the Methods section "Bone remodeling imaging and classification".   We agree and have included additional data measuring the calcium concentration in the HSPC reporter mice (MDS1 GFP , Christodoulou et al, 2020) in which the MDS1-driven GFP expression is not truncated by the expression of Flt3 (a gene associated with early differentiation). We found no difference between long-term HSCs and HSPCs in their local [Ca 2+ ]e (Fig R3 above, (Fig R4a, Suppl. Fig 9a in the  revised manuscript). In addition, we plotted the [Ca 2+ ]e against the distance to the endosteal bone fronts and again did not observe a spatial gradient (Fig R4b, Suppl. Fig. 3b in the revised manuscript) contrary to our expectation. Interestingly, we did observe a larger spread of Rhod5N/AF488 ratios close to the bone fronts, which is not due to measurement uncertainty since the signal-to-noise ratio (SNR) is usually higher near the endosteum (more superficial) than the deeper bone marrow regions. The reason for the variation needs to be investigated further but could be due to activities related to bone remodeling near the bone interface. Despite the larger spread in [Ca 2+ ]e close to the bone front, all HSCs were identified within 15 µm to the bone and were observed only in the high extracellular calcium regions. Minor point: 1) The number of HSCs analyzed in Figure 4C is unclear. Although the authors described that native HSCs were not found in the lowest Ca2+ region, I wonder if minor HSC population around the lowest Ca2+ region could not be found due to insufficient observed number of HSCs.

Response:
We thank the reviewer for pointing this out. We have clarified the HSC number and included more HSCs from n=7 to n=15 in the updated figure (Fig R3 in comment#6, also Fig 4c in the revised  manuscript). Typically, only 3-5 long-term HSCs could be found per calvarium, and we tried to include HSCs from all three cavity types (n=3, 7, 5 cells from D-, M-, R-type cavities, respectively, N=5 mice, Suppl. Figure 9c). In all cases HSCs were found near relatively high calcium regions. We have also included new data showing the [Ca 2+ ]e around individual HSPCs (n=30) and again found them to be in locations with relatively high [Ca 2+ ]e.

Reviewer #2 (Remarks to the Author):
Yeh et al present in their work an in vivo ratiometric fluorescence imaging method to measure (the very high) calcium levels in the extracellular space of calvarial bone marrow (interstitial space), taking into account the acidification state in different areas (related e.g. with bone resorption or bone formation). The main aim of developing the method is to elucidate which impact the extracellular calcium levels at various BM sites adjacent to bone have on the diversity of survival niches of early HSC (LT-HSC). Developing reliable, minimally invasive methods for mapping extracellular calcium in bone and bone marrow in vivo, in a spatialtemporal manner, is of highest relevance not only for the HSC/MSC research community and the cancer research community but also for the immunology community in general (immunological memory is just an example). Hence, I expect the approach presented by the Lin lab to be potentially of high impact after ruling out several issues regarding accuracy and reliability of the method, as discussed in the following.
The authors, from a lab having pioneered in vivo two-and three-photon microscopy of calvarial bone and bone marrow, chose for their calcium imaging approach Rhod-5N due to its high Kd, necessary to quantify interstitial calcium, also adjacent to bone. They coupled this red fluorescence calcium dye with the green emitting Alexa 488 dye to enable ratiometric measurements and correct the spectral signal ratios for pH value and effects of Rayleigh scattering of bone and marrow separately and developed an easy to use approach which I expect to find broad application in the bioscientific/biomedical community. The described normalization steps of the spectral signals is absolutely necessary, however, in order to ensure the reliability of the determined absolute calcium concentrations retrived by the presented method further validation is needed. Especially, validation using a different fluorescence method -fluorescence lifetime imaging -, which can be applied in vivo and is, in general, less affected by experimental circumstances would make the approach much stronger. While certainly not perfectly fitting, here are some suggestions: fluorescence lifetime imaging of CaG5N has been succesfully employed in measuring interstitial calcium levels

Response:
We thank the reviewer for the comment. As described in the response for Reviewer 1, comment #1, we have performed additional validation experiments using a calcium-independent dye, Rhodamine-B dextran (70kDa) and split the relatively broad emission spectrum of this single emitter into two (green and red) channels. The red-to-green (R/G) ratio should be invariant with depth and location in tissue. We showed that the apparent red shift with imaging depth can be corrected using the depth correction algorithm described in the manuscript. These results are shown in Fig R1 above  We have performed an additional validation experiment by thinning the bone with femtosecond laser-mediated ablation (Fig R1 e-f). We acquired 3D stacks of the same Rhodamine-B dextranlabeled bone marrow cavity (same field of view) just before and just after laser bone thinning. As shown in panel Fig R1 g-h above, the same dye exhibited different red shift depending on the thickness of the bone above it, attesting to the tissue-optics origin of the red shift. By applying the two-step correction, we are able to recover the same R/G ratio as the ratio determined for Rhodamine-B in vitro.
With respect to FLIM, while we agree that FLIM is generally considered a more robust method than ratiometric imaging, we do not think it is suitable in this case for multiple reasons articulated in response for Reviewer 1, comment #1 above. We have also added this in a new paragraph 9 of the Discussion.  2. Also for SNARF-1 such a full titration curve depicting the sensitivity throughout the pH dynamic range would strengthen the method.

Response:
We have included a full SNARF-1 calibration curve (Suppl Fig 13b, also Fig R6 below) that includes the 90% confidence intervals. 3. Except for one example, interstitial/extracellular calcium maps at only one single time-point have been acquired and used in the correction algorithm taking into account the different (mainly) Rayleigh scattering of the red (Rhod-5N) vs. green (Alexa 488) emitted fluorescence. However, heterogeneity will certainly occur also over time -this aspect needs to be taken into account when validating the correction algorithm. I expect that effects of different photobleaching behaviour -and, in general, photophysical behaviour -of the two dyes will change their spectral signal ratio and, thus, the determined calcium concentrations (as well as the pH values determined by SNARF1). Especially, as published by the lab, the oxygen levels (pO2) strongly vary throughout the bone marrow -this has an impact on the cells, but also some order of magnitude below in scale, on the dye molecules too: the fluorescence deplition as well as spectral shifts of excitation and fluorescence spectra (just as an example, Stockes / anti-Stockes shifts) are dramatically influenced by the depletion of the first triplet state of the dyes -how much this is populated due to inter-system crossing in Rhod-5N vs. Alexa 488 needs to be taken into account. Hence, the influence of local oxygen concentrations on the fluorescence signals ratio (and on calcium levels) needs to be verified for full accuracy. Of course, the same hold true for SNARF-1, since differently ionized forms of the same molecule are expected to have different energetical levels (not only ground and first singlet state, but also different triplet states, imposing a change in the inter-system crossing rate and possible effect of oxygen). We agree with the reviewer that the measured R/G ratio can vary not just in space but also in time. Indeed we performed both depth correction and clearance rate correction (to account for the changing Rhod-5N/AF488 ratio due to differences in their clearance kinetics), as shown in Suppl. The reviewers made a good point about the potential triplet state buildup in the low oxygen microenvironment of the bone marrow. To investigate the role of oxygen concentration, we measured the Rhod5N/AF488 ratio in vitro by varying oxygen concentrations in the calibration samples (via nitrogen bubbling). The absolute oxygen concentration was determined by a Unisense microelectrode, and the measurement was carried out in two different calcium concentrations. We showed that the ratios (Rhod5N/AF488) remained constant, independent of pO2 (Fig R7, also Suppl Fig. 12 in the revised manuscript), indicating that triplet state buildup is not a significant factor.
We have added these to paragraph 9 in the Discussion of the revised manuscript.  (Fig R8) are in close agreement with these numbers. Note that our attenuation lengths (30-40 µm) are closer to 1/µs than 1/µs' because we are imaging through a bone thickness of < 60 µm, which is not sufficient to reach the diffusion regime. We have added Fig R8 as Suppl. Fig. 1 a-b in the revised manuscript and added a discussion (Paragraph 7) comparing our measurements to previous publications of mean free paths in bone tissue. Another publication by Wang et al (Biomed Opt Express 2019, 10(4), 1905-1918) measured the Texas red (R) to fluorescein (G) signal ratio in mouse brain tissue excited at a single wavelength (three-photon excitation at 1450 nm). Normalized to the R/G ratio at the surface, the ratio increased to 1.34 at the depth of 600 µm, and to 1.43 at 800 µm. In comparison, our normalized R/G ratio for Rhod-5N and AF488 increased to ~1.2 at a depth of only 50 µm in the bone marrow (Fig 2b). The ratio would go up to ~4.5 if we extrapolated our measurement to 600 µm deep into the bone marrow. The stronger wavelength dependence can be explained by not just the higher scattering of the bone tissue, but also higher absorption due to the much higher vascular density in the bone marrow than in the brain. The blood volume fraction is estimated to be ~25% in the bone marrow [Kunisaki et al, Nature 2013, Spencer et al, Nature 2014] compared to ~3-5% in the brain [Leenders et al. Brain, 113,1990]. As a result, we expect a steeper increase of R/G ratio in the bone marrow compared to the brain.
5. Additionally, both scattering effects and effects of wave front distortions (lens effects of blood vessels, spherical aberrations, astigmatism) are not only depth-dependent but vary also within single tissue layers (in x and y). Whereas a quantification would be extremly tidious, at least estimating the impact of these effects would have on the accuracy of pH values and calcium levels is necessary. The accuracy will be impaired -the question is, if relevant differences between different regions in the bone marrow can still be detected, given the uncertainty caused by these effects. Response:

Attenuation coefficients Attenuation lengths
We agree with the reviewer that the wave front distortions could potentially impact the accuracy of the ratiometric measurement. However, since both the red and the green signals are excited by the same wavelength, the effect of wavefront distortion (degradation of the point spread function of the excitation beam) is mostly to reduce the laser intensity at the focus, resulting in a reduction in signal to noise ratio (SNR). We verified that the measured R/G ratios are largely independent of laser power (Fig R9, also Suppl. Fig 12b in the revised manuscript), except when the SNR falls very low (slight increase in the R/G ratio). The low SNR voxels are excluded in our analysis. This is now more clearly stated in the Methods section ("Image Processing").

Reviewer #3 (Remarks to the Author):
In this manuscript Yeh and colleagues report the development of an intravital imaging-based technique that allows to perform spatially resolved measurements of pH and calcium in the bone marrow (BM) microenvironment. The approach makes use of calcium sensitive ratiometric probes, and previously established, widely used multiphoton intravital imaging of calvarial BM tissue in mice, in which the authors are experts. Notably, the method requires the implementation of mathematical correction of the attenuation of fluorescence with imaging depth of the probes, which depends on the both the thickness of the bone as well as that of marrow tissue that light needs to go through. Using this technique, the authors are able to provide measurements of the interstitial pH and [Ca2+] in BM tissues and they report differences between endosteal regions depending on the metabolic state of the proximal bone surface. Finally, by using a reporter mouse of hematopoietic stem cells they also assess the pH and [Ca2+] in the immediate vicinity of HSCs, thus providing estimation of these parameters in the HSC niche. The BM is the primary site for hematopoiesis and hematopoietic stem cell maintenance. A critical question, which remains unresolved to date, is the specific cellular and molecular composition of the anatomical niches in which HSCs reside. Furthermore, to what extent different spatial compartments of the marrow differ in their physiological conditions is of great interest to the understanding of spatial compartmentalization in this tissue. Thus, the ability to simultaneously perform spatially resolved measurements of key parameters dictating cell fate, such as oxygen levels, [Ca2+] or pH, in situ and in vivo and in a non-invasive fashion, is of great relevance, technically very challenging and of high merit. I have some comments on specific points, which in my view could improve the manuscript 1. The main caveat to the study is acknowledged by the authors in the Discussion. The need to pair the calcium indicator probe with a reference dye requires that both dyes have the same biodistribution, or otherwise this could lead to inaccurate measurements depending on the tissue region imaged. While it could be assumed that both dyes used in the study may indeed not vary much in their biodistribution, this is difficult to ascertain at this point and casts doubt as to the accuracy of the data. Indeed, from the images in Supplemental video 3, it would seem as the dyes are not always evenly distributed. In the discussion, the authors propose an elegant way to circumvent this, the coupling of probes to low molecular weight dextrans, whose biodistributions would be equal. Given that dextran conjugation is relatively straightforward, why is this approach not tested here to ultimately confirm the validity of their technique? When possible it would be desirable to have this experiment done and the measurements repeated and compared to the dextran-free approach.
Response: Indeed, the biodistribution of the dye pair is critical, and as pointed out by the reviewer, conjugating both dyes to dextran would address this issue. However, we were advised by colleagues in chemistry that conjugation of Rhod-5N to dextran would alter its calcium response, as the same moiety for conjugation is also used for calcium binding. Instead, we packaged Rhod5N and AF488 with a fix stoichiometry into micelles (~125 nm hydrodynamic radius), so the packaged dyes had the same biodistribution, and were cleared at the same rate. We validated the ratiometric measurement using a strict comparison: the animal was injected with micelles first for obtaining R/G ratios. Once the micelles were cleared (~4 hours), we injected our original dye mixture (with two separate dyes) to obtain the R/G ratio of the same bone marrow cavity for comparison. In Fig R2 above (Figure 3e, Suppl. Fig. 3  ] underneath the osteoclast sealing zone (which is inaccessible to both the calcium sensor and the HSCs) and not in the bone marrow interstitial space. Thus, the calcium liberated from the bone resorption site is either not released as free calcium ions or is heavily buffered.
We did observe a larger spread of Rhod5N/AF488 ratios close to the endosteum, which is not due to measurement uncertainty since the signal-to-noise ratio (SNR) is usually higher near the endosteum (more superficial) than the deeper bone marrow regions (Fig. R4b). It is possible that the activities related to bone remodeling near the bone interface can give rise to the observed fluctuation in the [Ca 2+ ]e but further studies are needed to address the cause.
We have also analyzed the [Ca 2+ ] in the perivascular regions and found no difference between arterioles and sinusoidal blood vessels, or between peri-vascular regions vs. the endosteal zone (< 10 µm) (Fig. R10)  3. Similarly, are there local differences in pH between different regions of the BM? According to Figure 1f the pH values in the interstitium range from 6.8 to 7.4. Are these variations related to spatial location?
Response: We further analyzed the pH distribution of different regions of the bone marrow with respect to the bone fronts (Fig. R11, Suppl. Fig 3a in the revised manuscript). The results showed a slight but statistically significant decrease of pH (mean = 7.23 to 7.15) beyond the endosteal zone (< 10 µm). Larger data spread was observed within 10 µm to bone surface, potentially attributed to differences in the microenvironment required for osteoblast or osteoclast activation  Measured interstitial pH with respect to the distance to bone surface (N=3 animals, n=10 bone marrow cavities). Box and whiskers represent the median, 25 and 75 percentiles, and the 10-90% data range. Two-sided Mann-Whitney test. In the scatter plot, each dot represents an average ratio from a manually selected region (~3-cell radius) 4. In the discussion the authors mention that LT-HSCs were not found in areas with lowest [Ca2+] levels, however this trend is not quantified and shown in the Figure. Response: We have included additional quantification and description in the Result section (Fig 4c in the  revised manuscript and Fig R3 above). Based on the updated measurements of LT-HSCs, the calcium concentration near MFG cells span from 0.7-2.7 mM, with 10% and 90% percentile between 0.8 and 2.6 mM. In comparison, the interstitial calcium in vivo span from 0.1 -4.5 mM, with 10% and 90% percentile between 0.5 and 1.6 mM, supporting the observation that LT-HSCs are not found in low calcium regions in the interstitial space.
5. Along the same lines, the measurements of both pH and [Ca2+] in the vicinity of HSCs are interesting, but do not inform on whether the values for both parameters are exclusive or distinct for HSCs, or all cell types in the BM are exposed to similar conditions. The authors could use reporter mice for other cell types, for instance, B cells, T cells or neutrophils and perform similar measurements that can be used as a reference to understand this issue. We thank the reviewer for this comment and have included additional data measuring the calcium concentration in the HSPC reporter mice (MDS1 GFP , Christodoulou et al, 2020) in which the MDS1-driven GFP expression is not truncated by the expression of Flt3 (a gene associated with early differentiation). We found no difference between long-term HSCs and HSPCs in their local [Ca 2+ ]e (Fig R3 above, also Fig 4c in the revised manuscript). Both HSCs and the HSPCs reside in locations with higher [Ca 2+ ]e compared to the serum [Ca 2+ ] and to the overall [Ca 2+ ]e in the bone marrow. As hematopoietic cells downstream of the HSCs and HSPCs eventually populate the entire marrow, some of the more mature cells will move down the calcium gradient. Whether this is accomplished by specific subpopulations of hematopoietic cells remains to be determined.   Minor points: • The authors should provide details on the spatial resolution in all dimensions of these measurements of pH and [Ca2+] in the main text.
Response: We agree that high spatial resolution is essential for accurate segmentation of the interstitial space. We estimated the resolution (point spread function) by measuring the apparent width of the interstitial space (the narrow gap between cells where the fluorescence signals come from) in our images. Using this method, the transverse (x,y) resolution is estimated to be ~0.95 µm (Fig R14, also Suppl. Fig. 1g in the revised manuscript). The axial resolution is limited by the 3 µm step size along the z direction. The large step size is necessary because the z stacks need to be acquired in a time short compared to the rapid dye clearance. We have clarified this in the results. • In Figure 2a, it would be interesting to provide the SHG image to visualize the collagen signal of the areas marked as osteoids.